messages
listlengths
3
13
images
listlengths
1
1
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000496160.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000403394.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000381193.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000297517.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000509492.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000410808.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000351025.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000196028.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000114101.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000380395.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000207992.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000525615.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000087217.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000354761.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000006336.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000508200.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000517148.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000100347.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000214478.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000224695.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000402624.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000533240.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000072101.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000207455.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000081589.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000023630.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000038426.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000192575.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000496683.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000232755.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000382924.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000530433.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000073801.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000420475.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000055607.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000203611.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000391388.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000421623.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000569065.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000082980.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000285024.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000066389.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000397619.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000520808.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000566746.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000342702.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000579386.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000305422.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000044943.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000376785.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000319293.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000544933.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000298799.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000254821.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000322008.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000134747.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000461619.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000542484.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000111823.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000295753.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000345549.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000294937.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000337541.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000297964.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000449387.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000077874.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000484646.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000316481.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000050290.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000070356.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000032918.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000474075.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000311448.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000567287.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000126356.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000433472.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000388928.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000155029.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000495902.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000387887.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000094660.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000078384.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000238717.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000568000.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000419332.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000006107.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000307426.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000215984.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000180571.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000081845.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000191869.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000141874.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000151371.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000398426.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000516119.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000233703.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000310242.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000328810.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000006990.jpg" ]
[ { "role": "system", "content": "You are a helpful vision-language assistant. You should produce accurate, detailed, and grounded answers based on the image and the user's instructions. When given feedback, critique, or scores, revise your response to improve correctness, specificity, and completeness.", ...
[ "/outputs/image_base/coco/train2017/000000394539.jpg" ]