Dataset Viewer
id
stringlengths 29
32
| messages
listlengths 2
2
|
|---|---|
data_synthesis_sft_ticker_0_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $BYND with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$BYND - JPMorgan reels in expectations on Beyond Meat \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_1_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $CCL with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$CCL $RCL - Nomura points to bookings weakness at Carnival and Royal Caribbean \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_1_1 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $CCL with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$CCL $RCL $NCLH - Saudis take stake in Carnival Corp. - shares up 27% \", \"sentiment\": 1}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_1_2 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $CCL with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"Carnival $CCL shares climbed after a Saudi investment fund bought an 8.2% stake, valued at nearly $370 million, in\\u2026 \", \"sentiment\": 1}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_2_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $CX with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$CX - Cemex cut at Credit Suisse, J.P. Morgan on weak building outlook \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_3_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $ESS with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$ESS: BTIG Research cuts to Neutral \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_4_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $FNKO with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$FNKO - Funko slides after Piper Jaffray PT cut \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_5_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $FTI with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$FTI - TechnipFMC downgraded at Berenberg but called Top Pick at Deutsche Bank \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_6_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $GM with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$GM - GM loses a bull \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_6_1 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $GM with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$GM: Deutsche Bank cuts to Hold \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_6_2 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $GM with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$GM - Morgan Stanley sees GM doubling in bull case scenario \", \"sentiment\": 1}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_7_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $GTT with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$GTT: Cowen cuts to Market Perform\", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_7_1 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $GTT with neutral (2) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$GTT - GTT Communications appoints Steven Berns as CFO \", \"sentiment\": 2}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_8_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $HNHAF with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$HNHAF $HNHPD $AAPL - Trendforce cuts iPhone estimate after Foxconn delay \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_9_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $HOG with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$HOG - Moody's warns on Harley-Davidson \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_9_1 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $HOG with neutral (2) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"RECAP 12/10 Chatter:\\n$HOG + Vague Chatter\\n$IBM + Starboard\\n$FIT - $GOOGL Deal DOJ Probe\\n$TEN + $APO Powertrain Bid\\u2026 \", \"sentiment\": 2}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_10_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $HXL with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$HXL - Citing aero ties, Wells slashes PT on Hexcel \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_11_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $I with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$I - Intelsat cut to Market Perform at Raymond James \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_12_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $KRG with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$KRG: Compass Point cuts to Sell \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_13_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $LK with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$LK - Muddy Waters goes short Luckin Coffee \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_13_1 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $LK with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$LK - Citron and Muddy Waters at odds over Luckin Coffeee \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_13_2 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $LK with neutral (2) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$LK calls @BlackBox_Team \", \"sentiment\": 2}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_13_3 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $LK with neutral (2) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"Dear $LK Management:\\n\\nI've taken the liberty of prepping your response to today's Muddy Waters report from our \\\"res... \", \"sentiment\": 2}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_13_4 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $LK with neutral (2) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$LK - Luckin Coffee: It Make Sense That Shorts Sell At 4x Sales.. Read more:\\u00a0\\u00a0#finance #investing #stocks\", \"sentiment\": 2}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_14_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $MANT with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$MANT - ManTech downgraded ahead of difficult comps \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_15_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $MDCO with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$MDCO: Oppenheimer cuts to Perform\", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_15_1 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $MDCO with neutral (2) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$MDCO $NVS The Medicines Company Enters Into Definitive Agreement to be Acquired by Novartis AG for $9.7 Billion \", \"sentiment\": 2}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_15_2 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $MDCO with neutral (2) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$MDCO: The Medicines Co to be acquired by Novartis (NVS) for $85.00/share in cash \", \"sentiment\": 2}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_15_3 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $MDCO with neutral (2) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$MDCO \", \"sentiment\": 2}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_15_4 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $MDCO with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$MDCO (+22.6% pre) Novartis $NVS to Swallow Up The Medicines Company for $9.7 Billion - TheStreet\\n\\n\", \"sentiment\": 1}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_15_5 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $MDCO with neutral (2) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$MDCO $BCRX $AQST - Healthcare - Top 5 Gainers / Losers \", \"sentiment\": 2}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_16_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $MPLX with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$MPLX $MPC - MPLX cut at Credit Suisse on potential dilution from Marathon strategic review \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_17_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $MSGN with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$MSGN - Imperial downgrades MSG Networks amid sports-free airwaves \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_18_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $MTLS with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$MTLS - Piper hits the Materialise sidelines \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_19_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $NCBS with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$NCBS: Hovde Group cuts to Market Perform\", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_20_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $NFLX with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$NFLX - New Netflix bear steps out \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_20_1 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $NFLX with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"Wells Fargo Downgrades Netflix $NFLX to Underperform but sees as a takeover target. NFLX could get acquired: NFLX i\\u2026 \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_20_2 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $NFLX with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$NFLX - SunTrust expecting Netflix to post strong user growth \", \"sentiment\": 1}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_20_3 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $NFLX with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$NFLX (-0.7% pre) Wells Fargo Downgrades Netflix (NFLX) to Underperform on Free Cash Flow Concerns, Still Sees Acqu\\u2026 \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_21_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $SHOP with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$SHOP - Shopify loses a bull \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_22_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $STAY with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$STAY - Nomura Instinet loses confidence in Extended Stay America \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_23_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $TWLO with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$TWLO - Twilio gets Street-low target on virus risk \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_23_1 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $TWLO with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$TWLO: Twilio beats by $0.03, beats on revs; guides Q1 EPS below consensus, revs above consensus; guides FY20 EPS...\\u00a0\", \"sentiment\": 1}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_24_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $VCLT with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$VCLT $SPLB $IGLB - Guggenheim's Minerd sees more gloom ahead in stocks \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_25_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $VIAC with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$VIAC $VIACA $FOX - MoffettNathanson expects NFL TV costs to skyrocket \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_25_1 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $VIAC with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$VIAC: Barrington Research cuts to Mkt Perform \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_25_2 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $VIAC with neutral (2) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"ICYMI, @ViacomCBS is now on Nasdaq $VIAC! \\n\\n\\ud83d\\udc4b to\\n\\n@ParamountPics \\n@MTV \\n@Nickelodeon \\n@ComedyCentral \\n@BET\\u2026 \", \"sentiment\": 2}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_26_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $XLNX with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$XLNX - Mizuho cuts XLNX target on near-term headwinds \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_27_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $AEO with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"Downgrades 11/25: $AEO $COLM $DGII $EQT $ESS $EVRG $GTT $KRG $MANT $MDCO $NFLX $OMP $PEG $PRU $SBT $SITC $TAP $VIPS\", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_28_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $MLND with bearish (0) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"Downgrades 4/7: $MLND $MSGN $MTLS $O $OC $OVV $PEB $PK $PLYA $QEP $RHP $RPAI $SAIA $SDC $SHOP $SKY $SLG $SM $TWLO\\u2026 \", \"sentiment\": 0}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_29_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $ALTG with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$ALTG: Dougherty & Company starts at Buy\", \"sentiment\": 1}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_30_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $AMD with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$AMD - AMD's Navi shows strong adoption - BofA \", \"sentiment\": 1}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_30_1 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $AMD with neutral (2) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$AMD - Crash Protection For AMD. #economy #finance #business\", \"sentiment\": 2}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_31_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $ANCUF with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$ANCUF: BMO Capital Markets ups to Outperform\", \"sentiment\": 1}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_32_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $BHE with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$BHE: Lake Street starts at Buy\", \"sentiment\": 1}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_33_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $CNC with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$CNC *S&P Raises Centene Corp. Rtg To BBB- From BB+; Outlk Stable*\", \"sentiment\": 1}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_34_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $DE with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$DE - Citi gives big boost to Deere PT \", \"sentiment\": 1}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_35_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $EBAY with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$EBAY - Guggenheim sees favorable catalysts for eBay \", \"sentiment\": 1}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_35_1 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $EBAY with neutral (2) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$EBAY - EBay nears deal to sell StubHub - WSJ \", \"sentiment\": 2}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_35_2 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $EBAY with neutral (2) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$EBAY could disclose an agreement to sell StubHub to Viagogo for about $4 billion \", \"sentiment\": 2}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_35_3 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $EBAY with neutral (2) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$EBAY: eBay confirms deal to sell StubHub to viagogo for a purchase price of $4.05 billion in cash \", \"sentiment\": 2}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_35_4 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $EBAY with neutral (2) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"eBay $EBAY reportedly nears deal to sell StubHub for $4 billion\", \"sentiment\": 2}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_35_5 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $EBAY with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$EBAY - EBay +6% after selling StubHub \", \"sentiment\": 1}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_35_6 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $EBAY with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$EBAY (+4.5% pre) EBay Nears Deal to Sell StubHub to Viagogo for Roughly $4 Billion - WSJ\\n\\n\", \"sentiment\": 1}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_36_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $FCX with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$FCX - Freeport upgraded as Credit Suisse sees big buying opportunity \", \"sentiment\": 1}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_36_1 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $FCX with neutral (2) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$FCX: Freeport-McMoRan builds on this week's strength \", \"sentiment\": 2}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_37_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $FLXN with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$FLXN: BTIG Research starts at Buy \", \"sentiment\": 1}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_38_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $FNV with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$FNV - Franco-Nevada will not need more funding despite suspensions, analyst says \", \"sentiment\": 1}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_39_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $GBT with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$GBT: Cantor Fitzgerald resumes at Overweight\", \"sentiment\": 1}",
"role": "assistant"
}
]
|
data_synthesis_sft_ticker_40_0 | [
{
"content": "Generate one realistic financial news headline about stock tickers following real-world financial reporting for sentiment analysis.\n\nGuidelines for Generating Headlines:\n\n1. Format & Style:\n\t•\tHeadlines must be concise and mimic real financial news.\n\t•\tUse sentence case formatting (capitalize only the first word and proper nouns).\n\t•\tSome headlines should start with a stock ticker (e.g., $AAPL -), while others should begin with the company name or a broader market trend.\n\n2. Ticker Inclusion:\n\t•\tAt least one headline should include a stock ticker (e.g., $TSLA - or $NVDA -).\n\t•\tSome headlines should refer to companies by name instead of tickers (e.g., “Alphabet and Meta see price targets cut at Barclays”).\n\n3. Common Financial Themes:\n\nEnsure headlines reflect realistic financial news topics, including:\n\t•\tStock downgrades/upgrades\n\t•\tPrice target adjustments\n\t•\tMarket trends/economic outlook\n\t•\tCompany performance concerns\n\t•\tCompany news\n\t•\tCompany announcements\n\t•\tCompany events\n\n4. Source Attribution:\n\t•\tWhen relevant, mention an investment firm, analyst, or research group (e.g., Morgan Stanley, Barclays, Oppenheimer).\n\t•\tDo not fabricate research firms—use only well-known institutions.\n\n5. Sentiment Labeling:\n\nEach headline must be assigned a sentiment label based on its tone:\n\t•\tBearish (0): Indicates negative sentiment about a stock or market trend.\n\t•\tBullish (1): Indicates positive sentiment about a stock or market trend.\n\t•\tNeutral (2): Indicates neutral or informational tone.\n\nFormat:\nThe output should be formatted as a JSON instance that conforms to the JSON schema below.\n\nAs an example, for the schema {\"properties\": {\"foo\": {\"title\": \"Foo\", \"description\": \"a list of strings\", \"type\": \"array\", \"items\": {\"type\": \"string\"}}}, \"required\": [\"foo\"]}\nthe object {\"foo\": [\"bar\", \"baz\"]} is a well-formatted instance of the schema. The object {\"properties\": {\"foo\": [\"bar\", \"baz\"]}} is not well-formatted.\n\nHere is the output schema:\n```\n{\"properties\": {\"headline\": {\"description\": \"financial news headline\", \"title\": \"Headline\", \"type\": \"string\"}, \"sentiment\": {\"description\": \"sentiment (0=bearish, 1=bullish, 2=neutral)\", \"title\": \"Sentiment\", \"type\": \"string\"}}, \"required\": [\"headline\", \"sentiment\"]}\n```\n\nNow, generate one financial news headline about stock tickers: $GOOS with bullish (1) label following these guidelines. \nYour JSON output:",
"role": "user"
},
{
"content": "{\"headline\": \"$GOOS: Barclays starts at Overweight\", \"sentiment\": 1}",
"role": "assistant"
}
]
|
End of preview. Expand
in Data Studio
README.md exists but content is empty.
- Downloads last month
- 31