Update system prompt

README.md

@@ -112,10 +112,10 @@ It achieves the following results on the evaluation set:
 
 The model performs best in summarization tasks, specifically in English and maybe Chinese. The model provides reasoning ON/OFF via system prompt trigger, all instructions should be contained within the user prompt.
 
-Reasoning on example:
+Reasoning off example:
 ```json
 messages = [
-    {"role": "system", "content": "reasoning on"},
+    {"role": "system", "content": "reasoning off"},
     {"role": "user", "content": "Summarize the following into 5 bullet points, each with 20 words max.\n\nMarch 28 (Reuters) -..."}
 ]
 
@@ -123,10 +123,10 @@ messages = [
 - Elon Musk's xAI acquires X ...
 ```
 
-Reasoning off example:
+Reasoning on example:
 ```json
 messages = [
-    {"role": "system", "content": "reasoning off"},
+    {"role": "system", "content": "reasoning on"},
     {"role": "user", "content": "Summarize the following into 5 bullet points, each with 20 words max.\n\nMarch 28 (Reuters) -..."}
 ]