image_path stringlengths 26 138 | images images listlengths 1 1 | easyr1_prompt stringlengths 224 1.34k | bbox listlengths 4 4 | image_width int64 196 3.75k | image_height int64 224 2.21k | normalized_bbox listlengths 4 4 | instruction stringlengths 3 1.08k | conf_prompt stringlengths 1.71k 2.78k | no_conf_prompt stringlengths 634 1.71k |
|---|---|---|---|---|---|---|---|---|---|
element_grounding/10c63b5b7d80287543161b8248dc4fce.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
toggle extents and mouse position display | [
989,
1048,
1028,
1087
] | 1,932 | 1,092 | [
0.5119047619047619,
0.9597069597069597,
0.5320910973084886,
0.9954212454212454
] | toggle extents and mouse position display | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1192559.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "4+" | [
558,
941,
629,
989
] | 2,548 | 1,428 | [
0.21899529042386184,
0.6589635854341737,
0.2468602825745683,
0.6925770308123249
] | Click on "4+" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
slides_3840*2160/slide_64/original.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Select the center of the text box containing the phrase... | [
1164,
1121,
1178,
1135
] | 2,660 | 1,484 | [
0.43759398496240604,
0.7553908355795148,
0.44285714285714284,
0.7648247978436657
] | Select the center of the text box containing the phrase "Week 2." | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1276746.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Zoom" | [
2364,
1368,
2463,
1388
] | 2,548 | 1,428 | [
0.9277864992150706,
0.957983193277311,
0.966640502354788,
0.9719887955182073
] | Click on "Zoom" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1340448.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Properties" | [
450,
91,
532,
114
] | 2,548 | 1,428 | [
0.17660910518053374,
0.06372549019607843,
0.2087912087912088,
0.07983193277310924
] | Click on "Properties" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
doc_images/AmHHgw-Nep9dvztgVz7Wx_wWHjOh_1280x720_CHAR_8_21.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Given the following text: "Dolch Sight Words – Primer" ... | [
461,
338,
471,
348
] | 1,288 | 728 | [
0.3579192546583851,
0.4642857142857143,
0.3656832298136646,
0.47802197802197804
] | Given the following text: "Dolch Sight Words – Primer" , find the text in the document and click the 2th character "r" in the text. | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1174130.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Change Case" | [
471,
109,
507,
132
] | 1,288 | 728 | [
0.3656832298136646,
0.14972527472527472,
0.3936335403726708,
0.1813186813186813
] | Click on "Change Case" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1544363.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Hourly Forecast - Not Selected" | [
0,
226,
48,
275
] | 1,288 | 728 | [
0,
0.31043956043956045,
0.037267080745341616,
0.37774725274725274
] | Click on "Hourly Forecast - Not Selected" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1230149.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Cut" | [
50,
67,
96,
89
] | 2,548 | 1,428 | [
0.019623233908948195,
0.0469187675070028,
0.03767660910518053,
0.062324929971988796
] | Click on "Cut" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1281260.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Intense Reference" | [
471,
72,
543,
129
] | 1,288 | 728 | [
0.3656832298136646,
0.0989010989010989,
0.421583850931677,
0.1771978021978022
] | Click on "Intense Reference" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1355359.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Translator" | [
2507,
199,
2548,
238
] | 2,548 | 1,428 | [
0.9839089481946625,
0.13935574229691877,
1,
0.16666666666666666
] | Click on "Translator" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
professional-apps-grounding-images/prof_apps_0000817_frame_011678.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Search through blueprint components | [
84,
784,
111,
813
] | 2,660 | 1,484 | [
0.031578947368421054,
0.5283018867924528,
0.04172932330827068,
0.5478436657681941
] | Search through blueprint components | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
showui-web-images/showui_web_013777.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Tap on DENVER BRONCOS | [
158,
662,
479,
742
] | 1,624 | 924 | [
0.09729064039408868,
0.7164502164502164,
0.29495073891625617,
0.803030303030303
] | Tap on DENVER BRONCOS | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1487427.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "App available. Install GitHub" | [
892,
40,
932,
66
] | 1,288 | 728 | [
0.6925465838509317,
0.054945054945054944,
0.7236024844720497,
0.09065934065934066
] | Click on "App available. Install GitHub" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1381609.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Balloon FRVR" | [
745,
521,
830,
607
] | 1,288 | 728 | [
0.578416149068323,
0.7156593406593407,
0.6444099378881988,
0.8337912087912088
] | Click on "Balloon FRVR" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
slides_1920*1080/slide_28/original_bbox_1.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on the rotate handle of the bounding box containi... | [
613,
200,
623,
210
] | 1,932 | 1,092 | [
0.317287784679089,
0.18315018315018314,
0.322463768115942,
0.19230769230769232
] | Click on the rotate handle of the bounding box containing the image of tall skyscrapers reaching toward the sky with a bright sun and scattered clouds in the background. | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
professional-apps-grounding-images/prof_apps_0000257_frame_016625.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Hide or show all objects within this collection group | [
2257,
100,
2281,
125
] | 2,660 | 1,484 | [
0.8484962406015037,
0.0673854447439353,
0.8575187969924812,
0.08423180592991913
] | Hide or show all objects within this collection group | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
element_grounding/5d7d61084f84587147b0e3fd510234d4.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Identify the ui element that is nearest and located hor... | [
892,
49,
919,
83
] | 1,372 | 728 | [
0.6501457725947521,
0.0673076923076923,
0.6698250728862973,
0.11401098901098901
] | Identify the ui element that is nearest and located horizontally to the left of the "y axis". | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
autogui-images/autogui_2020707_4315461f.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
I'll give descriptions of a certain webpage element; pl... | [
79,
255,
185,
283
] | 896 | 504 | [
0.08816964285714286,
0.5059523809523809,
0.20647321428571427,
0.5615079365079365
] | I'll give descriptions of a certain webpage element; please identify where they are in this screenshot (with point). This element provides access to a specific section or archive of content within the website. | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
cell_right_bottom_corner/images/1920x1080_cell_F14.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Bottom right corner of F-14 | [
527,
439,
537,
449
] | 1,932 | 1,092 | [
0.2727743271221532,
0.40201465201465203,
0.2779503105590062,
0.41117216117216115
] | Bottom right corner of F-14 | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
element_grounding/f792af9614a678ad54513bd33021e7f8.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
EXPAND FORMATTING STYLE | [
1468,
112,
1503,
168
] | 1,680 | 1,036 | [
0.8738095238095238,
0.10810810810810811,
0.8946428571428572,
0.16216216216216217
] | EXPAND FORMATTING STYLE | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
professional-apps-grounding-images/prof_apps_0000526_frame_032994.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
See the active image being edited | [
1795,
1236,
2026,
1390
] | 2,660 | 1,484 | [
0.674812030075188,
0.8328840970350404,
0.7616541353383459,
0.9366576819407008
] | See the active image being edited | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1308979.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Language English (United States)" | [
167,
666,
374,
687
] | 1,288 | 728 | [
0.1296583850931677,
0.9148351648351648,
0.29037267080745344,
0.9436813186813187
] | Click on "Language English (United States)" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1347786.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Identity verification via Persona | LinkedIn ... | [
817,
0,
886,
39
] | 2,548 | 1,428 | [
0.3206436420722135,
0,
0.347723704866562,
0.0273109243697479
] | Click on "Identity verification via Persona | LinkedIn Help" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1536037.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "This site has coupons! Shopping in Microsoft ... | [
896,
40,
936,
66
] | 1,288 | 728 | [
0.6956521739130435,
0.054945054945054944,
0.7267080745341615,
0.09065934065934066
] | Click on "This site has coupons! Shopping in Microsoft Edge" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
slides_1280*720/slide_32/original_bbox_1.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Select the rotate handle of the text box containing the... | [
614,
296,
624,
306
] | 1,288 | 728 | [
0.4767080745341615,
0.4065934065934066,
0.484472049689441,
0.42032967032967034
] | Select the rotate handle of the text box containing the text "Your ability to communicate ef...". | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
showui-web-images/showui_web_001250.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Select 3 | [
491,
770,
525,
801
] | 1,708 | 980 | [
0.28747072599531615,
0.7857142857142857,
0.3073770491803279,
0.8173469387755102
] | Select 3 | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1175246.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "451 Like" | [
812,
868,
871,
896
] | 2,548 | 1,428 | [
0.31868131868131866,
0.6078431372549019,
0.34183673469387754,
0.6274509803921569
] | Click on "451 Like" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1210036.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Follow on Twitter" | [
396,
81,
409,
99
] | 1,288 | 728 | [
0.30745341614906835,
0.11126373626373626,
0.31754658385093165,
0.13598901098901098
] | Click on "Follow on Twitter" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
slides_1920*1080/slide_65/original_bbox_12.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Select the rotate handle of the text box containing the... | [
1229,
820,
1239,
830
] | 1,932 | 1,092 | [
0.6361283643892339,
0.7509157509157509,
0.6413043478260869,
0.76007326007326
] | Select the rotate handle of the text box containing the text "Main Street Market." | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
doc_images/AmHHgw-Nep9dvzk-PtsWxXoR51d4_1280x720_SPACE_52_52.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Given the following text: " the fastest way to refresh ... | [
516,
516,
526,
526
] | 1,288 | 728 | [
0.40062111801242234,
0.7087912087912088,
0.4083850931677019,
0.7225274725274725
] | Given the following text: " the fastest way to refresh your mind about details of the action " , find the text in the document and click the space between the continuous character "o" and "f" in the text. | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1446200.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Normal" | [
1003,
512,
1271,
544
] | 1,288 | 812 | [
0.7787267080745341,
0.6305418719211823,
0.9868012422360248,
0.6699507389162561
] | Click on "Normal" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1227788.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Arrange All" | [
810,
67,
899,
89
] | 2,548 | 1,428 | [
0.3178963893249607,
0.0469187675070028,
0.35282574568288855,
0.062324929971988796
] | Click on "Arrange All" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1411546.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "AutomationID: DynamicSearchBoxGleamImage" | [
1106,
755,
1184,
795
] | 1,288 | 812 | [
0.8586956521739131,
0.9298029556650246,
0.9192546583850931,
0.979064039408867
] | Click on "AutomationID: DynamicSearchBoxGleamImage" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
element_grounding/6bd5b01765a2fcc8d4e621b4c0e2359b.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Access audio mixer options | [
844,
960,
900,
1003
] | 1,932 | 1,092 | [
0.43685300207039335,
0.8791208791208791,
0.4658385093167702,
0.9184981684981685
] | Access audio mixer options | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
showui-web-images/showui_web_005774.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Choose Settings and more (Alt+F) | [
1742,
32,
1775,
63
] | 1,820 | 1,008 | [
0.9571428571428572,
0.031746031746031744,
0.9752747252747253,
0.0625
] | Choose Settings and more (Alt+F) | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1540817.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "More Options" | [
14,
104,
56,
141
] | 1,288 | 728 | [
0.010869565217391304,
0.14285714285714285,
0.043478260869565216,
0.1936813186813187
] | Click on "More Options" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
showui-web-images/showui_web_010089.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Tap on The Data Scientist’s Toolbox Course by Johns Hop... | [
462,
318,
795,
577
] | 2,352 | 1,344 | [
0.19642857142857142,
0.23660714285714285,
0.3380102040816326,
0.42931547619047616
] | Tap on The Data Scientist’s Toolbox Course by Johns Hopkins University, | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1478869.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Browser essentials" | [
1159,
38,
1199,
68
] | 1,288 | 728 | [
0.8998447204968945,
0.0521978021978022,
0.9309006211180124,
0.09340659340659341
] | Click on "Browser essentials" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
autogui-images/autogui_221914_6bf87dee.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Using the provided screenshot, I'll describe a webpage ... | [
31,
227,
90,
234
] | 224 | 364 | [
0.13839285714285715,
0.6236263736263736,
0.4017857142857143,
0.6428571428571429
] | Using the provided screenshot, I'll describe a webpage element for you to locate (with point). This element serves as a navigation tool, directing users to additional resources related to the blog post author or topic. | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
pc-agent-e-images/data/events/screenshot/cc4d_04f06604_2.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click the Settings option | [
969,
409,
1248,
437
] | 1,288 | 728 | [
0.7523291925465838,
0.5618131868131868,
0.968944099378882,
0.6002747252747253
] | Click the Settings option | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
showui-web-images/showui_web_005789.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Find and click Search Icon | [
801,
36,
820,
57
] | 1,008 | 560 | [
0.7946428571428571,
0.06428571428571428,
0.8134920634920635,
0.10178571428571428
] | Find and click Search Icon | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
slides_1280*720/slide_49/original.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Select the handle located in the bottom-left corner of ... | [
426,
403,
437,
413
] | 1,288 | 728 | [
0.3307453416149068,
0.5535714285714286,
0.3392857142857143,
0.5673076923076923
] | Select the handle located in the bottom-left corner of the text box containing the text "Thank you." | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1320546.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Clear Formatting" | [
570,
77,
592,
100
] | 1,288 | 728 | [
0.44254658385093165,
0.10576923076923077,
0.45962732919254656,
0.13736263736263737
] | Click on "Clear Formatting" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
slides_1280*720/slide_13/original_bbox_0.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Select the rotate handle of the text box containing the... | [
725,
254,
735,
264
] | 1,288 | 728 | [
0.562888198757764,
0.3489010989010989,
0.5706521739130435,
0.3626373626373626
] | Select the rotate handle of the text box containing the text "PROGRESS REPORT." | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
showui-web-images/showui_web_014860.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Navigate to Search Steam | [
2055,
362,
2102,
408
] | 2,632 | 1,484 | [
0.7807750759878419,
0.24393530997304583,
0.7986322188449848,
0.2749326145552561
] | Navigate to Search Steam | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1428221.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Visual Studio Code - 1 running window" | [
580,
1388,
629,
1428
] | 2,548 | 1,428 | [
0.22762951334379905,
0.9719887955182073,
0.2468602825745683,
1
] | Click on "Visual Studio Code - 1 running window" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
autogui-images/autogui_1171947_cef092c5.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
In this web page capture, please predict the positions ... | [
146,
144,
175,
159
] | 504 | 280 | [
0.2896825396825397,
0.5142857142857142,
0.3472222222222222,
0.5678571428571428
] | In this web page capture, please predict the positions of the element I describe (with point). This element serves as a navigation link, providing access to news articles or content related to Mexico on the webpage. Users can click on this link to explore news about Mexico within the 'Americas' section of the site. | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1180336.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Select" | [
1274,
111,
1342,
134
] | 2,548 | 1,428 | [
0.5,
0.07773109243697479,
0.5266875981161695,
0.0938375350140056
] | Click on "Select" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1314727.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Microsoft account | Privacy" | [
892,
0,
951,
35
] | 1,288 | 728 | [
0.6925465838509317,
0,
0.7383540372670807,
0.04807692307692308
] | Click on "Microsoft account | Privacy" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
element_grounding/599d61464a528ec23e179c1b2f8db6b8.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Identify the ui element that is nearest and located ver... | [
71,
1117,
262,
1152
] | 2,548 | 1,344 | [
0.027864992150706435,
0.8311011904761905,
0.10282574568288853,
0.8571428571428571
] | Identify the ui element that is nearest and located vertically below the "mastodon.social". | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1514580.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Google Account Help" | [
1136,
0,
1254,
41
] | 1,288 | 728 | [
0.8819875776397516,
0,
0.9736024844720497,
0.05631868131868132
] | Click on "Google Account Help" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
autogui-images/autogui_2240217_24936a0c.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
In the image above, I will give a series of description... | [
175,
92,
215,
108
] | 504 | 280 | [
0.3472222222222222,
0.32857142857142857,
0.42658730158730157,
0.38571428571428573
] | In the image above, I will give a series of descriptions of the element to be clicked. Please predict where you want to click (with point). This element triggers a sharing or saving functionality for users to share or save the current article they are viewing. | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1293941.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Teal" | [
703,
1316,
728,
1342
] | 2,548 | 1,428 | [
0.2759026687598116,
0.9215686274509803,
0.2857142857142857,
0.9397759103641457
] | Click on "Teal" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
slides_3840*2160/slide_33/original.png | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click the handle located at the bottom of the text box ... | [
793,
1193,
806,
1207
] | 2,660 | 1,484 | [
0.2981203007518797,
0.8039083557951483,
0.3030075187969925,
0.8133423180592992
] | Click the handle located at the bottom of the text box displaying the text "Learn to infuse energy into yo...." | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1344793.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Row up" | [
515,
136,
538,
168
] | 1,288 | 812 | [
0.3998447204968944,
0.16748768472906403,
0.41770186335403725,
0.20689655172413793
] | Click on "Row up" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... | |
gta-images/gta_1170028.jpg | You are an expert UI element locator. Given a GUI image and a user's element description, provide the coordinates of the specified element as a single (x,y) point. For elements with area, return the center point.
Output the coordinate pair exactly:
(x,y)
<image>
Click on "Hello, sign in Account & Lists" | [
1653,
75,
1786,
125
] | 2,548 | 1,428 | [
0.6487441130298273,
0.052521008403361345,
0.7009419152276295,
0.08753501400560224
] | Click on "Hello, sign in Account & Lists" | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
<analysis> (uncertainty, ambiguity, visibility issues) </analysis>
<confidence> 1-5 </confidence>
## Rules
- Identify the object that best ... | You are given an image and an instruction. Locate the referenced object and output its coordinates.
## Response Format
<think> (your reasoning) </think>
<answer> (x, y) </answer>
## Rules
- Identify the object that best matches the instruction.
- Output normalized coordinates (0–1000).
- If no suitable object is foun... |
End of preview. Expand in Data Studio
README.md exists but content is empty.
- Downloads last month
- 67