uid stringlengths 13 49 | category stringclasses 12
values | source stringclasses 37
values | task_family stringclasses 28
values | images listlengths 2 10 | question stringlengths 602 11.5k | answer stringlengths 340 2.85k | gold stringlengths 1 1.34k | teacher_model stringclasses 6
values | visual_audit stringclasses 2
values | second_teacher_agrees bool 2
classes |
|---|---|---|---|---|---|---|---|---|---|---|
imagecode_ed0f96830ff1_0 | Retrieval | imagecode | fine_grained_retrieval | [
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/video-storytelling-videowedding_cNoOzS2HZ5k-shot42_0/img0.jpg",
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/video-storytelling-videowedding_cNoOzS... | Given this detailed description:
There is a ninety degree arm coming out of the bottom left corner. The fist is right below the guy's head.
Which image provided best matches the description?
A. [Image 1]
B. [Image 2]
C. [Image 3]
D. [Image 4]
E. [Image 5]
F. [Image 6]
G. [Image 7]
H. [Image 8]
I. [Image 9]
J. [Image 1... | Understand: The question asks which image best shows a ninety degree arm coming out of the bottom left corner, with the fist right below the guy's head. The answer should be the image number that best matches this description.
Perceive:
- Image 1: The arm is extended outward but not at a ninety degree angle; fist is n... | 9 | gpt-4.1 | fail | null |
imagecode_16fa18a4101f_0 | Retrieval | imagecode | fine_grained_retrieval | [
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/MSR-VTT-videoTrainValVideo_video2028-shot5_1/img48.jpg",
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/MSR-VTT-videoTrainValVideo_video2028-shot5_1/... | [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7][Image 8][Image 9][Image 10]Given this detailed description:
the horse is standing to the right of the door, there is the smallest gap between his nose and the door frame, all legs are down
Which image provided best matches the description?
Reason step by ... | Understand: The question asks which image shows the horse standing to the right of the door, with all its legs down and the smallest gap between its nose and the door frame. The answer should be a single image number.
Perceive:
- Image 1: Horse is to the right of the door, all legs down, there is a visible gap between... | 3 | gpt-4.1 | fail | null |
jigsaw_d5b170b0e610_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_d5b170b0e610_0_hole.jpg",
"synthetic/jigsaw/jigsaw_d5b170b0e610_0_p0.jpg",
"synthetic/jigsaw/jigsaw_d5b170b0e610_0_p1.jpg",
"synthetic/jigsaw/jigsaw_d5b170b0e610_0_p2.jpg",
"synthetic/jigsaw/jigsaw_d5b170b0e610_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four patches (Images 2–5) matches the area covered by the grey rectangle in Image 1. The answer must be given as the option letter corresponding to the correct patch.
Perceive:
- Image 1: The grey rectangle covers part of the statue's torso and drapery, likely including folds... | C | gpt-4.1 | pass | true |
jigsaw_b35fbe0a5fce_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_b35fbe0a5fce_0_hole.jpg",
"synthetic/jigsaw/jigsaw_b35fbe0a5fce_0_p0.jpg",
"synthetic/jigsaw/jigsaw_b35fbe0a5fce_0_p1.jpg",
"synthetic/jigsaw/jigsaw_b35fbe0a5fce_0_p2.jpg",
"synthetic/jigsaw/jigsaw_b35fbe0a5fce_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the provided patches (A, B, C, D) fits exactly over the grey rectangle in Image 1, based on comparing the contents of this area with the patches. The answer must be the option letter corresponding to the patch that matches the grey rectangle’s content.
Perceive:
- Image 1: The gr... | B | gpt-4.1 | fail | null |
jigsaw_690c1eae0063_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_690c1eae0063_0_hole.jpg",
"synthetic/jigsaw/jigsaw_690c1eae0063_0_p0.jpg",
"synthetic/jigsaw/jigsaw_690c1eae0063_0_p1.jpg",
"synthetic/jigsaw/jigsaw_690c1eae0063_0_p2.jpg",
"synthetic/jigsaw/jigsaw_690c1eae0063_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four image patches (Images 2-5) was cut from the grey rectangle in Image 1. The answer must be an option letter corresponding to the image that matches the covered region.
Perceive:
- Image 1: The grey rectangle covers the chest and torso area of the woman wearing a pinkish s... | A | gpt-4.1 | pass | true |
VISION_81c82788b0c9_0 | Similarity | VISION | same_variety_truefalse | [
"VISION/full/images/2168.jpg",
"VISION/full/images/2169.jpg"
] | Are these two workpieces of the same kind? [Image 1][Image 2]
Choice list:
A: True
B: False
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the evidence in that image which bears on the qu... | Understand: The question asks whether the two workpieces shown in Image 1 and Image 2 are of the same kind, requiring a comparison based on their visible shapes, arrangement, and features.
Perceive:
- Image 1: Shows a set of circular structures with a prominent central depression, aligned in a regular grid; the edges ... | A | gpt-4.1 | pass | true |
VISION_cd7ae79021dc_0 | Similarity | VISION | same_variety_truefalse | [
"VISION/full/images/1664.jpg",
"VISION/full/images/1665.jpg"
] | Are these two workpieces of the same kind? [Image 1][Image 2]
Choice list:
A: True
B: False
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the evidence in that image which bears on the qu... | Understand: The question asks if the two workpieces (sets of printed circuit boards) in Image 1 and Image 2 are of the same kind, requiring a comparison of layout, components, and markings.
Perceive:
- Image 1: Shows a set of green printed circuit boards in a particular arrangement, with distinctive outlines, identical... | A | gpt-4.1 | pass | true |
imagecode_5567b5e41a3d_0 | Retrieval | imagecode | fine_grained_retrieval | [
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/MSR-VTT-videoTrainValVideo_video3099-shot1_0/img0.jpg",
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/MSR-VTT-videoTrainValVideo_video3099-shot1_0/i... | Given this detailed description:
Both arms are down to the side, no hands showing. The head is tilted back a little and the mouth is open. There is a blurry half of his face man in the bottom right corner
Which image provided best matches the description?[Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7][I... | Understand: The question asks which image best matches a description: both arms down, no hands showing, head tilted back, mouth open, and a blurry half-face man in the bottom right corner. The answer should be the single image number that fits all these traits.
Perceive:
- Image 1: Left arm is down, right hand holding... | 5 | gpt-4.1 | pass | true |
VISION_31702b06bd96_0 | Similarity | VISION | same_variety_truefalse | [
"VISION/full/images/1846.jpg",
"VISION/full/images/1847.jpg"
] | Are these two workpieces of the same kind? [Image 1][Image 2]
Choice list:
A: True
B: False
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the evidence in that image which bears on the qu... | Understand: The question asks whether the workpieces in Image 1 and Image 2 are of the same kind, requiring a comparison of their designs and layouts to determine similarity.
Perceive:
- Image 1: Contains a set of circuit boards with a specific layout, arrangement of components, labels, and text marking such as "EB85_P... | A | gpt-4.1 | fail | null |
nuscenes_530f35ddbd54_0 | Multi-view | nuscenes | multi_view_driving_qa | [
"nuscenes/full/images/45894.jpg",
"nuscenes/full/images/45895.jpg",
"nuscenes/full/images/45896.jpg",
"nuscenes/full/images/45897.jpg",
"nuscenes/full/images/45898.jpg",
"nuscenes/full/images/45899.jpg"
] | Upon receiving six photographs captured from various cameras on a street-view car, your responsibility is to provide accurate responses to questions about the scene. You must choose your answer from the Choice List. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6]Question: what is parked in the parking lot?
Reas... | Understand: The question asks for what is parked in the parking lot, so the answer must describe the type of vehicles seen in the parking lot across the images.
Perceive:
- Image 1: Many cars are parked in a lot surrounded by fencing.
- Image 2: Cars are parked along a sidewalk and in adjacent fenced parking area.
- Im... | cars | gpt-4.1 | pass | true |
jigsaw_428d6acae19a_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_428d6acae19a_0_hole.jpg",
"synthetic/jigsaw/jigsaw_428d6acae19a_0_p0.jpg",
"synthetic/jigsaw/jigsaw_428d6acae19a_0_p1.jpg",
"synthetic/jigsaw/jigsaw_428d6acae19a_0_p2.jpg",
"synthetic/jigsaw/jigsaw_428d6acae19a_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which patch (A, B, C, or D) exactly fills the grey rectangle region in Image 1. The answer must be a single option letter based on visual evidence from the images.
Perceive:
- Image 1: The grey rectangle is located over a section of the background, which looks like window blinds in a beig... | D | gpt-4.1 | pass | true |
VISION_b3e7bd3961b1_0 | Similarity | VISION | same_variety_truefalse | [
"VISION/full/images/4048.jpg",
"VISION/full/images/4049.jpg"
] | Are these two workpieces the same genre? [Image 1][Image 2]
Choice list:
A: True
B: False
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the evidence in that image which bears on the ques... | Understand: The question asks whether the two workpieces in Image 1 and Image 2 belong to the same genre, meaning if they are similar in type, function, or machining category.
Perceive:
- Image 1: Shows a metallic cylindrical cavity with visible machining marks, but the sides are smooth with no additional features or ... | A | gpt-4.1 | pass | false |
contrastive_caption_b5eecc3f004e_2 | Retrieval | contrastive_caption | image_text_matching | [
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/sharegpt4v/coco/train2017/000000167642.jpg",
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/sharegpt4v/coco/train2017/000000167644.jpg",
"contrastive... | [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7][Image 8]
Which image is the caption related to?
In the center of the image, a white plate with a blue rim is placed on a woven placemat. The plate is filled with a variety of food items, each adding a splash of color and texture to the scene. A serving of ... | Understand: The question asks which image matches the detailed description of a white plate with a blue rim, filled with pasta, grilled chicken, salad, and surrounded by a woven placemat, with two forks and glasses in the background. The answer must be the image number where these elements are visible.
Perceive:
- Ima... | 5 | gpt-4.1 | pass | true |
jigsaw_fcc52ae58084_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_fcc52ae58084_0_hole.jpg",
"synthetic/jigsaw/jigsaw_fcc52ae58084_0_p0.jpg",
"synthetic/jigsaw/jigsaw_fcc52ae58084_0_p1.jpg",
"synthetic/jigsaw/jigsaw_fcc52ae58084_0_p2.jpg",
"synthetic/jigsaw/jigsaw_fcc52ae58084_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four cropped patches best matches the region obscured by the grey rectangle in Image 1. The answer must be one of the options (A, B, C, D) corresponding to the four provided images.
Perceive:
- Image 1: The grey rectangle covers the upper center part of a girl's face (forehea... | C | gpt-4.1 | pass | true |
nuscenes_2cc8e3bb4696_0 | Multi-view | nuscenes | multi_view_driving_qa | [
"nuscenes/full/images/21594.jpg",
"nuscenes/full/images/21595.jpg",
"nuscenes/full/images/21596.jpg",
"nuscenes/full/images/21597.jpg",
"nuscenes/full/images/21598.jpg",
"nuscenes/full/images/21599.jpg"
] | With six images from multiple cameras on a street view car provided, your job is to answer questions pertaining to the scene. You must choose your answer from the Choice List. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6]Question: what color is the paint around the windows?
Reason step by step in this exact f... | Understand: The question asks for the color of the paint around the windows visible in the street view images, requiring direct observation of the window frames or trim color.
Perceive:
- Image 1: The windows have paint around them that is white.
- Image 2: Windows are not clearly visible or detailed enough to see trim... | white | gpt-4.1 | pass | true |
VISION_451e51aafca4_0 | Similarity | VISION | same_variety_truefalse | [
"VISION/full/images/15308.jpg",
"VISION/full/images/15309.jpg",
"VISION/full/images/15310.jpg",
"VISION/full/images/15311.jpg"
] | Are the damages on both workpieces of the same type? [Image 1][Image 2][Image 3][Image 4]
Choice list:
A: True
B: False
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the evidence in that... | Understand: The question asks whether the damages seen in the close-ups (Image 2 and Image 4) are of the same type on the workpieces shown in the larger boards (Image 1 and Image 3). The answer must be "True" if both damaged areas show the same type of damage, or "False" if they do not.
Perceive:
- Image 1: Shows a la... | A | gpt-4.1 | fail | null |
jigsaw_0c1e8f211913_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_0c1e8f211913_0_hole.jpg",
"synthetic/jigsaw/jigsaw_0c1e8f211913_0_p0.jpg",
"synthetic/jigsaw/jigsaw_0c1e8f211913_0_p1.jpg",
"synthetic/jigsaw/jigsaw_0c1e8f211913_0_p2.jpg",
"synthetic/jigsaw/jigsaw_0c1e8f211913_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four patches was cut from the region covered by the grey rectangle in Image 1. The answer should be the option letter corresponding to the patch that matches the covered area.
Perceive:
- Image 1: The grey rectangle covers a section of mushrooms in the lower center-left part.... | D | gpt-4.1 | pass | true |
contrastive_caption_d150f6940484_3 | Retrieval | contrastive_caption | image_text_matching | [
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/LaionGPT4V-LIVIS_images/30179.jpg",
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/LaionGPT4V-LIVIS_images/30180.jpg",
"contrastive_caption/0.0.0/c55... | [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7]
Which image is the caption describing?
a baby's hand reaching out from a crib, with two teddy bears placed alongside it. A pacifier is also present near the baby's hand. The largest teddy bear is located on the top left corner of the image, partially outsi... | Understand: The question asks for the image that shows a baby's hand reaching out from a crib, accompanied by teddy bears and a pacifier, with a peaceful mood. The answer must be the image number.
Perceive:
- Image 1: Shows a person kite surfing on the ocean at sunset.
- Image 2: Depicts a decorated cake with eggs and... | 6 | gpt-4.1 | pass | true |
jigsaw_4421c94df05e_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_4421c94df05e_0_hole.jpg",
"synthetic/jigsaw/jigsaw_4421c94df05e_0_p0.jpg",
"synthetic/jigsaw/jigsaw_4421c94df05e_0_p1.jpg",
"synthetic/jigsaw/jigsaw_4421c94df05e_0_p2.jpg",
"synthetic/jigsaw/jigsaw_4421c94df05e_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which patch (from options A-D, referencing images 2-5) was cut from the region covered by the grey rectangle in Image 1. The answer must be a single option letter.
Perceive:
- Image 1: Shows a man sitting indoors, with a grey rectangle covering a section near his hands. The area under the... | C | gpt-4.1 | pass | true |
jigsaw_63fc7bdd532e_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_63fc7bdd532e_0_hole.jpg",
"synthetic/jigsaw/jigsaw_63fc7bdd532e_0_p0.jpg",
"synthetic/jigsaw/jigsaw_63fc7bdd532e_0_p1.jpg",
"synthetic/jigsaw/jigsaw_63fc7bdd532e_0_p2.jpg",
"synthetic/jigsaw/jigsaw_63fc7bdd532e_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four patches (images 2 to 5) best matches the region currently covered by the grey rectangle in Image 1. The answer must be a single letter (A, B, C, or D), corresponding to the image that fits the covered area.
Perceive:
- Image 1: The grey rectangle covers the upper middle ... | C | gpt-4.1 | pass | true |
contrastive_caption_90693d1596aa_0 | Retrieval | contrastive_caption | image_text_matching | [
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/sharegpt4v/coco/train2017/000000180613.jpg",
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/sharegpt4v/coco/train2017/000000180623.jpg"
] | Which image is the caption related to?
In the given image, we see a picturesque scene of a beach where three young men are standing with their surfboards. The man on the left is holding a vibrant yellow surfboard that features a striking black lightning bolt design. The man in the middle is holding a bold red surfboard... | Understand: The question asks which image matches the detailed caption describing three men with surfboards on a beach, ocean in the background, "Mystic Wave" sign, and tropical vegetation. The answer must be the image number.
Perceive:
- Image 1: Shows three young men with surfboards, a large blue surfboard with "Myst... | 1 | gpt-4.1 | pass | true |
ScanQA_a778d53b0d69_0 | Multi-view | ScanQA | multi_view_3d_qa | [
"scannet_frames_25k/scene0368_00/color/000000.jpg",
"scannet_frames_25k/scene0368_00/color/000200.jpg",
"scannet_frames_25k/scene0368_00/color/000400.jpg",
"scannet_frames_25k/scene0368_00/color/000600.jpg",
"scannet_frames_25k/scene0368_00/color/000800.jpg",
"scannet_frames_25k/scene0368_00/color/001000.... | This is a 3D scenario. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7][Image 8]
What color board is to the left of the sofa chair?
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", givin... | Understand: The question asks about the color of the board to the left of the sofa chair in this 3D scenario. The answer must be a color.
Perceive:
- Image 1: Shows part of a sofa chair but no board visible to the left.
- Image 2: No sofa chair or board visible.
- Image 3: Shows a sofa chair, but no board to the left.... | white whiteboard | gpt-4.1 | fail | null |
VISION_f54aedf19a2c_0 | Similarity | VISION | same_variety_truefalse | [
"VISION/full/images/4624.jpg",
"VISION/full/images/4625.jpg"
] | Are these two workpieces the same type? [Image 1][Image 2]
Choice list:
A: True
B: False
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the evidence in that image which bears on the quest... | Understand: The question asks whether the two workpieces shown in Image 1 and Image 2 are of the same type, requiring a comparison of relevant visible features such as shape, size, and arrangement.
Perceive:
- Image 1: Shows circular objects arranged in a regular grid pattern; each object appears to have a similar size... | A | gpt-4.1 | pass | true |
contrastive_caption_4c2ce44050c2_0 | Retrieval | contrastive_caption | image_text_matching | [
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/sharegpt4v/coco/train2017/000000011156.jpg",
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/sharegpt4v/coco/train2017/000000011158.jpg",
"contrastive... | [Image 1] [Image 2] [Image 3] [Image 4] [Image 5] Which image is the caption connected to?
The image depicts a pair of **red-handled scissors** held in a person's hand against a **blue-green background** with white floral designs. The scissors are slightly open, and the word "**FISKARS**" is visible on one of the blade... | Understand: The question asks which image matches the detailed caption describing a pair of red-handled scissors, held in a hand, against a blue-green floral background with the word "FISKARS" visible.
Perceive:
- Image 1: Shows various items including electronics, a wallet, coins, and a bag; no scissors, hand, or flo... | 4 | gpt-4.1 | pass | true |
contrastive_caption_45b8c15ea1e5_1 | Retrieval | contrastive_caption | image_text_matching | [
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/LaionGPT4V-LIVIS_images/23790.jpg",
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/LaionGPT4V-LIVIS_images/23791.jpg",
"contrastive_caption/0.0.0/c55... | [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6]
What captures your attention in the the secondary image
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the evidence in that image wh... | Understand: The question is asking which element is most attention-grabbing in the secondary image among the provided set. The answer must be the image number only.
Perceive:
- Image 1: Two giraffes interacting in a grassy enclosure.
- Image 2: Three giraffes near a stack of logs in a wooded area.
- Image 3: Three gir... | 3 | gpt-4.1 | pass | false |
ScanQA_b7879b672760_0 | Multi-view | ScanQA | multi_view_3d_qa | [
"scannet_frames_25k/scene0048_00/color/000000.jpg",
"scannet_frames_25k/scene0048_00/color/000200.jpg",
"scannet_frames_25k/scene0048_00/color/000500.jpg",
"scannet_frames_25k/scene0048_00/color/000700.jpg",
"scannet_frames_25k/scene0048_00/color/000900.jpg",
"scannet_frames_25k/scene0048_00/color/001100.... | This is a 3D scenario. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7][Image 8]
What is the blue backpack on?
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the evidence i... | Understand: The question asks for the surface or object on which the blue backpack is placed, based on evidence from the images.
Perceive:
- Image 1: No blue backpack visible.
- Image 2: Blue backpack is on a light-colored round table.
- Image 3: Blue backpack is on a light-colored round table, seen closer.
- Image 4:... | top of round table | gpt-4.1 | pass | false |
contrastive_caption_d8bb24ace3c9_3 | Retrieval | contrastive_caption | image_text_matching | [
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/sharegpt4v/coco/train2017/000000086210.jpg",
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/sharegpt4v/coco/train2017/000000086213.jpg",
"contrastive... | [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7][Image 8]
Which image is the caption referring to?
In the heart of a vibrant restaurant, three individuals share a moment captured in time. The leftmost person, engrossed in the culinary delights before them, is poised with a silver fork mid-air, ready to s... | Understand: The question describes a scene in a vibrant restaurant with three people at a table: one eating with a fork, one using a cell phone, and one smiling at the camera. The table has white plates, silver utensils, a glass with a lime wedge, and colorful paintings in the background. The answer must identify the i... | 6 | gpt-4.1 | pass | true |
nuscenes_a5c72e5c3585_0 | Multi-view | nuscenes | multi_view_driving_qa | [
"nuscenes/full/images/13242.jpg",
"nuscenes/full/images/13243.jpg",
"nuscenes/full/images/13244.jpg",
"nuscenes/full/images/13245.jpg",
"nuscenes/full/images/13246.jpg",
"nuscenes/full/images/13247.jpg"
] | Presented with six images from various cameras on a street view car, your responsibility is to answer inquiries about the scene. You must choose your answer from the Choice List. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6]Question: what does the red and white bar lead to?
Reason step by step in this exact f... | Understand: The question asks where the red and white bar (typically a parking gate barrier) leads to, based on evidence visible in the images. The answer must be a location or area.
Perceive:
- Image 1: The red and white bar is part of a parking lot entrance with signs for parking and available lots; leads to rows of ... | parking | gpt-4.1 | pass | true |
jigsaw_9e06e98a5eca_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_9e06e98a5eca_0_hole.jpg",
"synthetic/jigsaw/jigsaw_9e06e98a5eca_0_p0.jpg",
"synthetic/jigsaw/jigsaw_9e06e98a5eca_0_p1.jpg",
"synthetic/jigsaw/jigsaw_9e06e98a5eca_0_p2.jpg",
"synthetic/jigsaw/jigsaw_9e06e98a5eca_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four image patches (options A-D, corresponding to Images 2-5) matches the covered region in Image 1. The answer must be one option letter.
Perceive:
- Image 1: The grey rectangle covers a region near the upper middle of the scene, beneath the shelter, above the fire, and over... | D | gpt-4.1 | pass | true |
jigsaw_3165f10671db_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_3165f10671db_0_hole.jpg",
"synthetic/jigsaw/jigsaw_3165f10671db_0_p0.jpg",
"synthetic/jigsaw/jigsaw_3165f10671db_0_p1.jpg",
"synthetic/jigsaw/jigsaw_3165f10671db_0_p2.jpg",
"synthetic/jigsaw/jigsaw_3165f10671db_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four patches (Images 2-5) was cut from the region covered by the grey rectangle in Image 1. The answer must be one option letter (A, B, C, or D) corresponding to the matching image.
Perceive:
- Image 1: The grey rectangle is placed over a region of beige floor, slightly refle... | B | gpt-4.1 | fail | null |
jigsaw_2c92a1dcd9aa_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_2c92a1dcd9aa_0_hole.jpg",
"synthetic/jigsaw/jigsaw_2c92a1dcd9aa_0_p0.jpg",
"synthetic/jigsaw/jigsaw_2c92a1dcd9aa_0_p1.jpg",
"synthetic/jigsaw/jigsaw_2c92a1dcd9aa_0_p2.jpg",
"synthetic/jigsaw/jigsaw_2c92a1dcd9aa_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four patch images matches the region covered by the grey rectangle in Image 1. The answer must be in the form of a single option letter (A, B, C, or D).
Perceive:
- Image 1: The grey rectangle covers an area near the horse's head, specifically the region where the mane and ea... | B | gpt-4.1 | pass | true |
jigsaw_c8f181d31b9d_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_c8f181d31b9d_0_hole.jpg",
"synthetic/jigsaw/jigsaw_c8f181d31b9d_0_p0.jpg",
"synthetic/jigsaw/jigsaw_c8f181d31b9d_0_p1.jpg",
"synthetic/jigsaw/jigsaw_c8f181d31b9d_0_p2.jpg",
"synthetic/jigsaw/jigsaw_c8f181d31b9d_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four image patches (Images 2–5) was cut from the grey rectangle covering part of the scene in Image 1. The answer must be a single option letter (A, B, C, or D) corresponding to the correct patch.
Perceive:
- Image 1: The grey rectangle covers part of a girl's face and upper ... | A | gpt-4.1 | fail | null |
jigsaw_3ebc3c179360_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_3ebc3c179360_0_hole.jpg",
"synthetic/jigsaw/jigsaw_3ebc3c179360_0_p0.jpg",
"synthetic/jigsaw/jigsaw_3ebc3c179360_0_p1.jpg",
"synthetic/jigsaw/jigsaw_3ebc3c179360_0_p2.jpg",
"synthetic/jigsaw/jigsaw_3ebc3c179360_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which patch among Images 2-5 corresponds to the area covered by the grey rectangle in Image 1. The answer must be a single option letter (A, B, C, or D).
Perceive:
- Image 1: The grey rectangle covers an area above the bird, primarily in the background, which consists mostly of branches a... | B | gpt-4.1 | pass | true |
imagecode_f76126f53c10_0 | Retrieval | imagecode | fine_grained_retrieval | [
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/MSR-VTT-videoTrainValVideo_video5326-shot1_1/img64.jpg",
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/MSR-VTT-videoTrainValVideo_video5326-shot1_1/... | Given this detailed description:
The tip of the knife is just barely behind the person's knuckle.
Which image provided best matches the description?
A. [Image 1]
B. [Image 2]
C. [Image 3]
D. [Image 4]
E. [Image 5]
F. [Image 6]
G. [Image 7]
H. [Image 8]
I. [Image 9]
J. [Image 10]
Reason step by step in this exact forma... | Understand: The question asks which image shows the tip of the knife just barely behind the person's knuckle, requiring careful observation of the knife's position relative to the knuckle in each image.
Perceive:
- Image 1: The knife tip is behind the knuckle, but there is a noticeable distance.
- Image 2: The knife ti... | 3 | gpt-4.1 | pass | false |
iconqa_7e27d52becb7_0 | Similarity | iconqa | diagram_choice | [
"iconqa/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/iconqa_data/iconqa/train/choose_img/85139/image.png",
"iconqa/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/iconqa_data/iconqa/train/choose_img/85139/choice_0.png",
"iconqa/0.0.0/c55... | [Image 1]Which is shaped like a sphere?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the evidence in that image which bears on the qu... | Understand: The question asks which object in Image 1 is shaped like a sphere, with answer options mapped to Images 2, 3, 4, and 5. The answer must be a letter corresponding to the correct option.
Perceive:
- Image 1: The first object from the left is a mirror disco ball, which is spherical. The other objects are not ... | C | gpt-4.1 | pass | true |
contrastive_caption_1a90bc0e7a26_1 | Retrieval | contrastive_caption | image_text_matching | [
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/sharegpt4v/coco/train2017/000000058765.jpg",
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/sharegpt4v/coco/train2017/000000058766.jpg",
"contrastive... | [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7]
Which image do you think the caption belongs to?
In the heart of a barren landscape, a light brown cow with majestic horns is the center of attention. The cow is indulging in a feast of colorful fruits and vegetables, scattered on a piece of paper that lie... | Understand: The question asks which image matches a detailed caption of a cow with majestic horns eating colorful fruits and vegetables on a piece of paper in a barren, sandy landscape, accompanied by a small black bird. The answer must be the image number only.
Perceive:
- Image 1: Contains a plate with food items, n... | 5 | gpt-4.1 | pass | true |
VISION_ad009e2b94c9_0 | Similarity | VISION | same_variety_truefalse | [
"VISION/full/images/4124.jpg",
"VISION/full/images/4125.jpg"
] | Are these two workpieces the same family? [Image 1][Image 2]
Choice list:
A: True
B: False
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the evidence in that image which bears on the que... | Understand: The question asks if the two workpieces in Image 1 and Image 2 are from the same family, meaning they should have similar shapes, structures, and features.
Perceive:
- Image 1: Shows a curved white surface with two diagonal grooves, minor dirt/marks on the upper groove.
- Image 2: Shows a curved white surfa... | A | gpt-4.1 | pass | true |
nuscenes_d06c25b6a5ec_0 | Multi-view | nuscenes | multi_view_driving_qa | [
"nuscenes/full/images/31392.jpg",
"nuscenes/full/images/31393.jpg",
"nuscenes/full/images/31394.jpg",
"nuscenes/full/images/31395.jpg",
"nuscenes/full/images/31396.jpg",
"nuscenes/full/images/31397.jpg"
] | Given a collection of six images taken from different cameras on a street view car, your task is to respond to the questions about the scenario. You must choose your answer from the Choice List. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6]Question: what color is the cab of the truck?
Reason step by step in t... | Understand: The question asks for the color of the cab of the truck, which is the front part where the driver sits. The answer must be a color.
Perceive:
- Image 1: The cab of the truck is clearly visible and it is blue.
- Image 2: The truck is visible but the cab color can't be determined since it is largely obscured ... | blue | gpt-4.1 | pass | true |
jigsaw_64ca5482a105_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_64ca5482a105_0_hole.jpg",
"synthetic/jigsaw/jigsaw_64ca5482a105_0_p0.jpg",
"synthetic/jigsaw/jigsaw_64ca5482a105_0_p1.jpg",
"synthetic/jigsaw/jigsaw_64ca5482a105_0_p2.jpg",
"synthetic/jigsaw/jigsaw_64ca5482a105_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which patch (from Images 2-5) matches the region covered by the grey rectangle in Image 1. The answer must be the option letter corresponding to the patch that fits the covered area.
Perceive:
- Image 1: The grey rectangle covers the center of the puppy's face, likely over the nose, fur, a... | B | gpt-4.1 | pass | true |
VISION_6daef37ea0d7_0 | Similarity | VISION | same_variety_truefalse | [
"VISION/full/images/1346.jpg",
"VISION/full/images/1347.jpg"
] | Are these two workpieces the same family? [Image 1][Image 2]
Choice list:
A: True
B: False
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the evidence in that image which bears on the que... | Understand: The question asks whether the two workpieces in Image 1 and Image 2 belong to the same family, likely meaning they share similar structural and visual characteristics. The answer must be one from the provided choice list.
Perceive:
- Image 1: Shows a grid of circular wells or pockets, all with similar size... | A | gpt-4.1 | pass | true |
ordering_0215427dc948_0 | Ordering | synthetic_ordering_AESOP | sequence_ordering | [
"AESOP/full/images/16033.jpg",
"AESOP/full/images/16032.jpg",
"AESOP/full/images/16034.jpg"
] | The 3 images are the panels of a short story, shown in a scrambled order.
Which option gives the correct order, first to last?
A. Image 2 -> Image 1 -> Image 3
B. Image 2 -> Image 3 -> Image 1
C. Image 3 -> Image 1 -> Image 2
D. Image 1 -> Image 3 -> Image 2
Reason step by step in this exact format:
Understand: state ... | Understand: The question asks for the logical order of the images to form a coherent short story, from first to last, based on their visual content.
Perceive:
- Image 1: A table is set for two, one person is present, apparently preparing or waiting.
- Image 2: The same person is in a kitchen, cooking with pots and pan... | A | gpt-4.1 | pass | true |
ScanQA_640b9c04cede_0 | Multi-view | ScanQA | multi_view_3d_qa | [
"scannet_frames_25k/scene0590_00/color/000000.jpg",
"scannet_frames_25k/scene0590_00/color/000400.jpg",
"scannet_frames_25k/scene0590_00/color/000800.jpg",
"scannet_frames_25k/scene0590_00/color/001200.jpg",
"scannet_frames_25k/scene0590_00/color/001600.jpg",
"scannet_frames_25k/scene0590_00/color/002000.... | This is a 3D scenario. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7][Image 8]
The door is tan and what shape?
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the evidence... | Understand: The question asks about the color and shape of the door in the scenario, which must be directly visually identified from the images.
Perceive:
- Image 1: The door is visible, and it is tan in color with a rectangular shape.
- Image 2: The same door is visible from a different angle, confirming its tan color... | rectangular | gpt-4.1 | pass | true |
MIT-States_StateCoherence_296e2173bdf6_0 | Similarity | MIT-States_StateCoherence | state_coherence_tf | [
"MIT-States_StateCoherence/full/images/4628.jpg",
"MIT-States_StateCoherence/full/images/4629.jpg",
"MIT-States_StateCoherence/full/images/4630.jpg",
"MIT-States_StateCoherence/full/images/4631.jpg"
] | I have two sets of pictures that depict an object in its opposite state. Can you tell me if the states of these two sets of pictures are the same? You must choose your answer from the Choice List. [Image 1][Image 2][Image 3][Image 4]
Choice list:
A: True
B: False
Reason step by step in this exact format:
Understand: s... | Understand: The question asks whether two sets of images in opposite states share the same kind of opposition (e.g., ripe/unripe, assembled/disassembled, etc.). The answer must be either A (True) or B (False).
Perceive:
- Image 1: Shows green, unripe fruit hanging from a tree.
- Image 2: Shows a ripe, red berry.
- Ima... | B | gpt-4.1 | pass | false |
jigsaw_7aa21b1a1992_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_7aa21b1a1992_0_hole.jpg",
"synthetic/jigsaw/jigsaw_7aa21b1a1992_0_p0.jpg",
"synthetic/jigsaw/jigsaw_7aa21b1a1992_0_p1.jpg",
"synthetic/jigsaw/jigsaw_7aa21b1a1992_0_p2.jpg",
"synthetic/jigsaw/jigsaw_7aa21b1a1992_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four patches (Image 2, 3, 4, 5) fits the grey rectangle in Image 1. The answer must be one of the option letters A, B, C, or D, corresponding to Images 2-5.
Perceive:
- Image 1: The grey rectangle covers the lower middle section of a boat with at least two people rowing. The ... | C | gpt-4.1 | fail | null |
ordering_174eafdb1732_0 | Ordering | synthetic_ordering_RecipeQA_ImageCoherence | sequence_ordering | [
"RecipeQA_ImageCoherence/full/images/44635.jpg",
"RecipeQA_ImageCoherence/full/images/44632.jpg",
"RecipeQA_ImageCoherence/full/images/44633.jpg",
"RecipeQA_ImageCoherence/full/images/44634.jpg"
] | The 4 images are the steps of a recipe, shown in a scrambled order.
Which option gives the correct order, first to last?
A. Image 3 -> Image 4 -> Image 1 -> Image 2
B. Image 2 -> Image 3 -> Image 1 -> Image 4
C. Image 4 -> Image 3 -> Image 2 -> Image 1
D. Image 2 -> Image 3 -> Image 4 -> Image 1
Reason step by step in... | Understand: The question asks for the correct chronological order of the four images showing steps of a recipe, from first to last. The answer must be a single option letter.
Perceive:
- Image 1: Shows chopped vegetables cooking in a pot, likely an early preparation step.
- Image 2: Shows a muffin tin with some dough ... | D | gpt-4.1 | pass | true |
jigsaw_5d88bf126981_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_5d88bf126981_0_hole.jpg",
"synthetic/jigsaw/jigsaw_5d88bf126981_0_p0.jpg",
"synthetic/jigsaw/jigsaw_5d88bf126981_0_p1.jpg",
"synthetic/jigsaw/jigsaw_5d88bf126981_0_p2.jpg",
"synthetic/jigsaw/jigsaw_5d88bf126981_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which patch (from Images 2-5) exactly fills the grey rectangle in Image 1, with each patch corresponding to an option (A-D). The answer must be a single option letter.
Perceive:
- Image 1: The grey rectangle covers the area in front of the child, roughly in the middle lower part of the ph... | D | gpt-4.1 | pass | true |
imagecode_2216698d25c8_0 | Retrieval | imagecode | fine_grained_retrieval | [
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/YouCook-video0007-shot1_10/img1928.jpg",
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/YouCook-video0007-shot1_10/img1932.jpg",
"imagecode/0.0.0/c... | Given this detailed description:
A non-blurry mixer sits on the left side of the bowl.
Which image provided best matches the description?
A. [Image 1]
B. [Image 2]
C. [Image 3]
D. [Image 4]
E. [Image 5]
F. [Image 6]
G. [Image 7]
H. [Image 8]
I. [Image 9]
J. [Image 10]
Reason step by step in this exact format:
Understa... | Understand: The question asks which image shows a non-blurry mixer positioned on the left side of the bowl, requiring identification of the image that best fits this description.
Perceive:
- Image 1: The mixer is on the left side but is blurry due to motion.
- Image 2: The mixer is on the left side but is still blurry.... | 3 | gpt-4.1 | pass | false |
contrastive_caption_7ae5ff5c7fc9_1 | Retrieval | contrastive_caption | image_text_matching | [
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/sharegpt4v/llava/llava_pretrain/images/00021/000211113.jpg",
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/sharegpt4v/llava/llava_pretrain/images/0002... | [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6]
Which image is the caption related to?
The image presents a playful and imaginative scene set in a residential area during daylight. Dominating the scene is a whimsical tree-like structure, ingeniously crafted from a stack of flower pots. The pots, painted in vibra... | Understand: The question asks which image matches the description of a whimsical tree-like structure made from colorful flower pots, with a striped small tree on top, set in a suburban area, and text referring to "Dr Seuss Tree" and "Janette's Saga".
Perceive:
- Image 1: Shows a stack of colorful flower pots, a small s... | 1 | gpt-4.1 | fail | null |
nuscenes_6e4173549a73_0 | Multi-view | nuscenes | multi_view_driving_qa | [
"nuscenes/full/images/38562.jpg",
"nuscenes/full/images/38563.jpg",
"nuscenes/full/images/38564.jpg",
"nuscenes/full/images/38565.jpg",
"nuscenes/full/images/38566.jpg",
"nuscenes/full/images/38567.jpg"
] | Presented with six images from various cameras on a street view car, your responsibility is to answer inquiries about the scene. You must choose your answer from the Choice List. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6]Question: what is the color of the student's shirts on the bus ad?
Reason step by step... | Understand: The question asks for the color of the student's shirts depicted in the bus advertisement. The answer should be a specific color.
Perceive:
- Image 1: No evidence relevant to student's shirts or bus ad.
- Image 2: Partial view of the bus ad, student shirts visible but not clear.
- Image 3: Clear view of the... | white | gpt-4.1 | fail | null |
ordering_14c744e4e9cc_0 | Ordering | synthetic_ordering_RecipeQA_ImageCoherence | sequence_ordering | [
"RecipeQA_ImageCoherence/full/images/689.jpg",
"RecipeQA_ImageCoherence/full/images/691.jpg",
"RecipeQA_ImageCoherence/full/images/688.jpg",
"RecipeQA_ImageCoherence/full/images/690.jpg"
] | The 4 images are the steps of a recipe, shown in a scrambled order.
Which option gives the correct order, first to last?
A. Image 1 -> Image 2 -> Image 4 -> Image 3
B. Image 3 -> Image 4 -> Image 1 -> Image 2
C. Image 3 -> Image 1 -> Image 4 -> Image 2
D. Image 4 -> Image 3 -> Image 2 -> Image 1
Reason step by step in... | Understand: The question asks for the correct chronological order of the four images as steps in a recipe, from first to last, based on visual evidence.
Perceive:
- Image 1: Pouring a liquid (possibly milk) into a pot, likely an early cooking step.
- Image 2: Metalworking equipment, not related to cooking; seems out of... | C | gpt-4.1 | pass | false |
nuscenes_442bc2215b41_0 | Multi-view | nuscenes | multi_view_driving_qa | [
"nuscenes/full/images/29808.jpg",
"nuscenes/full/images/29809.jpg",
"nuscenes/full/images/29810.jpg",
"nuscenes/full/images/29811.jpg",
"nuscenes/full/images/29812.jpg",
"nuscenes/full/images/29813.jpg"
] | Upon viewing six pictures captured from various cameras on a street view vehicle, your role is to answer questions about the presented scene. You must choose your answer from the Choice List. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6]Question: is this during the daytime?
Reason step by step in this exact f... | Understand: The question asks whether these images were taken during the daytime; the answer must be one from the Choice List, specifically referencing the presence of daylight.
Perceive:
- Image 1: Bright, natural light through glass, outdoor greenery well illuminated.
- Image 2: Street, buildings, and vehicles are c... | yes | gpt-4.1 | pass | true |
jigsaw_07a584ed2e5d_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_07a584ed2e5d_0_hole.jpg",
"synthetic/jigsaw/jigsaw_07a584ed2e5d_0_p0.jpg",
"synthetic/jigsaw/jigsaw_07a584ed2e5d_0_p1.jpg",
"synthetic/jigsaw/jigsaw_07a584ed2e5d_0_p2.jpg",
"synthetic/jigsaw/jigsaw_07a584ed2e5d_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four patches (A, B, C, D) fits exactly into the grey-covered area in Image 1. The answer must be the letter corresponding to the image most visually consistent with the covered region.
Perceive:
- Image 1: The grey rectangle is over a portion where there are pale birds (pelic... | D | gpt-4.1 | pass | true |
ordering_e1b99ddb3401_0 | Ordering | synthetic_ordering_AESOP | sequence_ordering | [
"AESOP/full/images/19983.jpg",
"AESOP/full/images/19985.jpg",
"AESOP/full/images/19984.jpg"
] | The 3 images are the panels of a short story, shown in a scrambled order.
Which option gives the correct order, first to last?
A. Image 3 -> Image 1 -> Image 2
B. Image 3 -> Image 2 -> Image 1
C. Image 1 -> Image 3 -> Image 2
D. Image 2 -> Image 1 -> Image 3
Reason step by step in this exact format:
Understand: state ... | Understand: The question asks for the correct chronological order of the three images as a short story. The answer must be an option (A, B, C, or D) representing the sequence from first to last.
Perceive:
- Image 1: The boy is sitting at a table with two books (one yellow, one red), a stool is occupied, and a blue boo... | C | gpt-4.1 | pass | false |
contrastive_caption_709454b2f53b_0 | Retrieval | contrastive_caption | image_text_matching | [
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/sharegpt4v/coco/train2017/000000191310.jpg",
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/sharegpt4v/coco/train2017/000000191314.jpg"
] | [Image 1] [Image 2] Which image do you think the caption belongs to?
In the heart of a serene neighborhood, a young girl is immersed in her own world. She's dressed in a vibrant purple jacket and contrasting pink pants, adding a splash of color to the otherwise ordinary day. Her attention is completely absorbed by the ... | Understand: The question asks which image matches a detailed caption describing a young girl in vibrant clothing, a scooter, a quiet street, and a suburban, tree-lined setting. The answer must be the image number.
Perceive:
- Image 1: Shows a young girl in a purple jacket and pink pants, walking with a scooter on a qu... | 1 | gpt-4.1 | fail | null |
jigsaw_0eea40898ea0_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_0eea40898ea0_0_hole.jpg",
"synthetic/jigsaw/jigsaw_0eea40898ea0_0_p0.jpg",
"synthetic/jigsaw/jigsaw_0eea40898ea0_0_p1.jpg",
"synthetic/jigsaw/jigsaw_0eea40898ea0_0_p2.jpg",
"synthetic/jigsaw/jigsaw_0eea40898ea0_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four patches (Images 2-5) exactly matches the grey-covered region of Image 1. The answer must be one option letter (A-D).
Perceive:
- Image 1: The grey rectangle covers the chest area of the person wearing an orange hoodie with large white text.
- Image 2: Shows the chest are... | A | gpt-4.1 | pass | true |
VISION_585d48f70fa6_0 | Similarity | VISION | same_variety_truefalse | [
"VISION/full/images/19120.jpg",
"VISION/full/images/19121.jpg",
"VISION/full/images/19122.jpg",
"VISION/full/images/19123.jpg"
] | Are the damages to the following two identical workpieces of the same type? [Image 1][Image 2][Image 3][Image 4]
Choice list:
A: True
B: False
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving on... | Understand: The question asks whether the damages to two identical workpieces, as shown in Images 2 and 4, are of the same type. The answer must be either True (A) or False (B).
Perceive:
- Image 1: Shows multiple assembled circuit boards; does not focus on any damage details.
- Image 2: Shows a close-up of a circuit b... | A | gpt-4.1 | pass | true |
jigsaw_1c597acd0dc7_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_1c597acd0dc7_0_hole.jpg",
"synthetic/jigsaw/jigsaw_1c597acd0dc7_0_p0.jpg",
"synthetic/jigsaw/jigsaw_1c597acd0dc7_0_p1.jpg",
"synthetic/jigsaw/jigsaw_1c597acd0dc7_0_p2.jpg",
"synthetic/jigsaw/jigsaw_1c597acd0dc7_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of four patches (Images 2-5) matches the region covered by the grey rectangle in Image 1. The answer must be a single option letter corresponding to the image that fits the covered region.
Perceive:
- Image 1: The grey rectangle covers the middle-upper part of the scene, above the h... | A | gpt-4.1 | pass | true |
jigsaw_7a242943710e_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_7a242943710e_0_hole.jpg",
"synthetic/jigsaw/jigsaw_7a242943710e_0_p0.jpg",
"synthetic/jigsaw/jigsaw_7a242943710e_0_p1.jpg",
"synthetic/jigsaw/jigsaw_7a242943710e_0_p2.jpg",
"synthetic/jigsaw/jigsaw_7a242943710e_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four patches (Images 2–5) fits the grey-covered region on the mushroom in Image 1. The answer must be a single option letter (A, B, C, or D).
Perceive:
- Image 1: The grey rectangle covers part of the lower mushroom body, which is white, rounded, and shows shallow grooves and... | C | gpt-4.1 | pass | true |
imagecode_afc469e82925_0 | Retrieval | imagecode | fine_grained_retrieval | [
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/open-images-2150_3500de53b7a80395/img0.jpg",
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/open-images-2150_3500de53b7a80395/img1.jpg",
"imagecode... | Given this detailed description:
A girl in black and white strip dress with a a white lace shirt over it.
Which image provided best matches the description?[Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7][Image 8][Image 9][Image 10]
Reason step by step in this exact format:
Understand: state what the qu... | Understand: The question asks which image best matches the description of a girl wearing a black and white striped dress with a white lace shirt over it. The answer must be the image number that fits this description.
Perceive:
- Image 1: Shows a person in a black and white striped dress, but no white lace shirt is vi... | 9 | gpt-4.1 | pass | true |
ScanQA_7b457c278d44_0 | Multi-view | ScanQA | multi_view_3d_qa | [
"scannet_frames_25k/scene0675_00/color/000000.jpg",
"scannet_frames_25k/scene0675_00/color/000200.jpg",
"scannet_frames_25k/scene0675_00/color/000400.jpg",
"scannet_frames_25k/scene0675_00/color/000600.jpg",
"scannet_frames_25k/scene0675_00/color/000800.jpg",
"scannet_frames_25k/scene0675_00/color/000300.... | This is a 3D scenario. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7][Image 8]
What is sitting under the desk in the right corner?
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", givi... | Understand: The question asks what object is sitting underneath the desk located in the right corner of the room. The answer must be a brief identification of the object.
Perceive:
- Image 1: The right corner desk has a chair positioned at it, with the chair clearly visible.
- Image 2: Desk not visible; only shows bed... | black office chair | gpt-4.1 | pass | true |
ScanQA_b30d9fab36db_0 | Multi-view | ScanQA | multi_view_3d_qa | [
"scannet_frames_25k/scene0510_00/color/000000.jpg",
"scannet_frames_25k/scene0510_00/color/000300.jpg",
"scannet_frames_25k/scene0510_00/color/000600.jpg",
"scannet_frames_25k/scene0510_00/color/000900.jpg",
"scannet_frames_25k/scene0510_00/color/001200.jpg",
"scannet_frames_25k/scene0510_00/color/001500.... | This is a 3D scenario. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7][Image 8]
What color is the warning sign underneath the shelf?
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giv... | Understand: The question asks for the color of the warning sign beneath the shelf in these bathroom images; the answer must be a brief color description.
Perceive:
- Image 1: No warning sign visible.
- Image 2: No warning sign visible.
- Image 3: A yellow warning sign is visible in the reflection, but not under a shelf... | yellow | gpt-4.1 | fail | null |
imagecode_6fde5ea56fb2_0 | Retrieval | imagecode | fine_grained_retrieval | [
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/MSR-VTT-videoTrainValVideo_video1873-shot1_15/img494.jpg",
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/MSR-VTT-videoTrainValVideo_video1873-shot1_... | [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7][Image 8][Image 9][Image 10]Given a detailed description, retrieve the target image among 10 minimally contrastive images
Description:
The dog is pointing straight towards the camera. It's tail is sticking straight up in line with its body.
Reason step by ... | Understand: The question asks for the image where the dog is facing directly towards the camera, with its tail sticking straight up and in line with its body. The answer format is a single digit corresponding to the image number.
Perceive:
- Image 1: The dog’s body is mostly submerged, and its head is not fully visibl... | 10 | gpt-4.1 | fail | null |
ScanQA_72e0429661c0_0 | Multi-view | ScanQA | multi_view_3d_qa | [
"scannet_frames_25k/scene0705_00/color/000000.jpg",
"scannet_frames_25k/scene0705_00/color/000200.jpg",
"scannet_frames_25k/scene0705_00/color/000400.jpg",
"scannet_frames_25k/scene0705_00/color/000600.jpg",
"scannet_frames_25k/scene0705_00/color/000800.jpg",
"scannet_frames_25k/scene0705_00/color/001000.... | This is a 3D scenario. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7][Image 8]
On what side of the room is the whiteboard located?
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", givi... | Understand: The question asks for the specific side of the room where the whiteboard is located, based on the 3D scenario shown in the images.
Perceive:
- Image 1: The whiteboard is visible on the right wall next to the TV screen.
- Image 2: The whiteboard is on the right wall, adjacent to the table and chairs.
- Imag... | right | gpt-4.1 | pass | true |
contrastive_caption_0dbb34eda42d_1 | Retrieval | contrastive_caption | image_text_matching | [
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/sharegpt4v/coco/train2017/000000193261.jpg",
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/sharegpt4v/coco/train2017/000000193271.jpg",
"contrastive... | [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7]
Which image is the caption describing?
The image captures a moment of serene beauty, featuring a seagull in mid-flight over a body of water. The seagull, with its wings spread wide, is the central figure in this composition. Its wingspan nearly matches the... | Understand: The question asks which image fits a caption describing a serene scene of a seagull in mid-flight over tranquil blue water, featuring a white seagull with black wingtips, and the seagull as the central figure.
Perceive:
- Image 1: Shows an urban street with buildings, no water or bird present.
- Image 2: Sh... | 7 | gpt-4.1 | pass | true |
imagecode_9d8ff48a03d2_0 | Retrieval | imagecode | fine_grained_retrieval | [
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/open-images-1386_33189438ee33c4f4/img0.jpg",
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/open-images-1386_33189438ee33c4f4/img1.jpg",
"imagecode... | Given this detailed description:
the guy in white shirt and a woman wearing a read dress talking to each other while sitting in front at the conference. Also her dog is laying next to her.
Which image provided best matches the description?
A. [Image 1]
B. [Image 2]
C. [Image 3]
D. [Image 4]
E. [Image 5]
F. [Image 6]
G.... | Understand: The question asks which image shows a man in a white shirt and a woman in a red dress talking to each other while seated in front at a conference, with her dog laying next to her. The answer must be the index number of the image that best matches this scene.
Perceive:
- Image 1: No clear evidence of a man i... | 3 | gpt-4.1 | pass | true |
contrastive_caption_bb0c2a37569f_2 | Retrieval | contrastive_caption | image_text_matching | [
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/LaionGPT4V-LIVIS_images/29204.jpg",
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/LaionGPT4V-LIVIS_images/29205.jpg",
"contrastive_caption/0.0.0/c55... | [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6]
Which image is the caption related to?
a man wearing protective gear, including a helmet and ear protection, working on a large fallen tree in a wooded or forested area. Nearby, there is a red fire hydrant. The man appears to be cutting or trimming the fallen tree ... | Understand: The question is asking which image shows a man equipped with protective gear (helmet, ear protection) working on a large fallen tree, with a red fire hydrant nearby, presumably using a chainsaw in a wooded area. The answer should be the image number that matches this description.
Perceive:
- Image 1: A man... | 1 | gpt-4.1 | pass | true |
jigsaw_d2fe747b6f1e_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_d2fe747b6f1e_0_hole.jpg",
"synthetic/jigsaw/jigsaw_d2fe747b6f1e_0_p0.jpg",
"synthetic/jigsaw/jigsaw_d2fe747b6f1e_0_p1.jpg",
"synthetic/jigsaw/jigsaw_d2fe747b6f1e_0_p2.jpg",
"synthetic/jigsaw/jigsaw_d2fe747b6f1e_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four patches (Images 2–5) corresponds to the region covered by the grey rectangle in Image 1. The answer must be a single option (A, B, C, or D) representing the image that matches the covered region.
Perceive:
- Image 1: Shows a rabbit with a grey rectangle covering a centra... | A | gpt-4.1 | pass | true |
imagecode_83df5138ae7e_0 | Retrieval | imagecode | fine_grained_retrieval | [
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/open-images-1126_1f6d5d4adb244e4e/img0.jpg",
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/open-images-1126_1f6d5d4adb244e4e/img1.jpg",
"imagecode... | Given this detailed description:
The camera shot is from the bottom up at a camel. The camel is looking towards the left side of the picture. There is a sunrise or sunset in the background.
Which image provided best matches the description?
A. [Image 1]
B. [Image 2]
C. [Image 3]
D. [Image 4]
E. [Image 5]
F. [Image 6]
G... | Understand: The question is asking which image matches a description of a bottom-up camera shot of a camel looking left, with a sunrise or sunset in the background. The answer must identify the correct image number matching these details.
Perceive:
- Image 1: A camel statue shot from the side, with daylight and mountai... | 9 | gpt-4.1 | pass | true |
jigsaw_4f3a40eb0a44_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_4f3a40eb0a44_0_hole.jpg",
"synthetic/jigsaw/jigsaw_4f3a40eb0a44_0_p0.jpg",
"synthetic/jigsaw/jigsaw_4f3a40eb0a44_0_p1.jpg",
"synthetic/jigsaw/jigsaw_4f3a40eb0a44_0_p2.jpg",
"synthetic/jigsaw/jigsaw_4f3a40eb0a44_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four patches (Image 2-5) matches the region covered by the gray rectangle in Image 1. The answer must be the option letter that corresponds to the correct patch.
Perceive:
- Image 1: The gray rectangle covers part of the middle/right side of the scene, specifically above and ... | C | gpt-4.1 | pass | true |
VISION_3dadd892f408_0 | Similarity | VISION | same_variety_truefalse | [
"VISION/full/images/26300.jpg",
"VISION/full/images/26301.jpg",
"VISION/full/images/26302.jpg",
"VISION/full/images/26303.jpg"
] | Do both workpieces have identical damages? [Image 1][Image 2][Image 3][Image 4]
Choice list:
A: True
B: False
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the evidence in that image whi... | Understand: The question asks whether both workpieces shown in the images have identical damages, requiring comparison of surface imperfections or damages across the relevant images.
Perceive:
- Image 1: Shows a cylindrical metal surface with some visible scratches or machining marks, but no significant gouges or deep ... | B | gpt-4.1 | pass | true |
nuscenes_d3156d1d9fb9_0 | Multi-view | nuscenes | multi_view_driving_qa | [
"nuscenes/full/images/32226.jpg",
"nuscenes/full/images/32227.jpg",
"nuscenes/full/images/32228.jpg",
"nuscenes/full/images/32229.jpg",
"nuscenes/full/images/32230.jpg",
"nuscenes/full/images/32231.jpg"
] | Upon receiving six photographs captured from various cameras on a street-view car, your responsibility is to provide accurate responses to questions about the scene. You must choose your answer from the Choice List. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6]Question: does this place have any apple trees in ... | Understand: The question asks whether there are any apple trees in the location shown across all six images. The answer must be "yes" or "no," based on direct visual evidence of apple trees.
Perceive:
- Image 1: Shows a truck, some trees, none identifiable as apple trees.
- Image 2: Shows a street and landscaped area ... | no | gpt-4.1 | pass | true |
jigsaw_b640bbdedfcb_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_b640bbdedfcb_0_hole.jpg",
"synthetic/jigsaw/jigsaw_b640bbdedfcb_0_p0.jpg",
"synthetic/jigsaw/jigsaw_b640bbdedfcb_0_p1.jpg",
"synthetic/jigsaw/jigsaw_b640bbdedfcb_0_p2.jpg",
"synthetic/jigsaw/jigsaw_b640bbdedfcb_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four patches (from Images 2–5) fits the grey rectangle covering part of Image 1. The answer must be the option letter corresponding to the correct image.
Perceive:
- Image 1: The grey rectangle is located towards the upper-left center of the snow scene, which has a border betw... | A | gpt-4.1 | fail | null |
MIT-States_StateCoherence_a622799d6cce_0 | Similarity | MIT-States_StateCoherence | state_coherence_tf | [
"MIT-States_StateCoherence/full/images/3056.jpg",
"MIT-States_StateCoherence/full/images/3057.jpg",
"MIT-States_StateCoherence/full/images/3058.jpg",
"MIT-States_StateCoherence/full/images/3059.jpg"
] | I have two sets of pictures that depict an object in its opposite state. Can you tell me if the states of these two sets of pictures are identical? You must choose your answer from the Choice List. [Image 1][Image 2][Image 3][Image 4]
Choice list:
A: True
B: False
Reason step by step in this exact format:
Understand: ... | Understand: The question asks whether there are two sets of images, each depicting an object in two opposite states, and if the states shown in the two sets are identical. The answer should be chosen from the given Choice List.
Perceive:
- Image 1: Two identical owl-shaped objects facing away from each other, possibly ... | B | gpt-4.1 | fail | null |
jigsaw_16cb5bc83c68_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_16cb5bc83c68_0_hole.jpg",
"synthetic/jigsaw/jigsaw_16cb5bc83c68_0_p0.jpg",
"synthetic/jigsaw/jigsaw_16cb5bc83c68_0_p1.jpg",
"synthetic/jigsaw/jigsaw_16cb5bc83c68_0_p2.jpg",
"synthetic/jigsaw/jigsaw_16cb5bc83c68_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four patches (Images 2–5) was cut from the region covered by the grey rectangle in Image 1. The answer must be one option letter (A, B, C, or D).
Perceive:
- Image 1: The grey rectangle covers an area where the bird's body meets the branch, with some small red berries and gre... | C | gpt-4.1 | pass | false |
jigsaw_f0681f03d03c_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_f0681f03d03c_0_hole.jpg",
"synthetic/jigsaw/jigsaw_f0681f03d03c_0_p0.jpg",
"synthetic/jigsaw/jigsaw_f0681f03d03c_0_p1.jpg",
"synthetic/jigsaw/jigsaw_f0681f03d03c_0_p2.jpg",
"synthetic/jigsaw/jigsaw_f0681f03d03c_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four patches (Images 2-5) was cut from the region covered by the grey rectangle in Image 1. The answer must be a single option letter (A, B, C, or D).
Perceive:
- Image 1: The grey rectangle lies over the upper center-right of the image, partly covering a child in a green shi... | D | gpt-4.1 | fail | null |
nuscenes_7b87a93b5b85_0 | Multi-view | nuscenes | multi_view_driving_qa | [
"nuscenes/full/images/11790.jpg",
"nuscenes/full/images/11791.jpg",
"nuscenes/full/images/11792.jpg",
"nuscenes/full/images/11793.jpg",
"nuscenes/full/images/11794.jpg",
"nuscenes/full/images/11795.jpg"
] | With six images, each from a different camera on a street-view car, your assignment is to answer questions about the scene. You must choose your answer from the Choice List. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6]Question: what is the color of the buildings?
Reason step by step in this exact format:
Und... | Understand: The question asks for the color of the buildings visible in the images. The answer must be a brief description of the color(s).
Perceive:
- Image 1: The building is white.
- Image 2: A portion of the same white building as in Image 1 is visible.
- Image 3: A white building is clearly visible.
- Image 4: Th... | white | gpt-4.1 | pass | true |
ScanQA_c903b1d6edc9_0 | Multi-view | ScanQA | multi_view_3d_qa | [
"scannet_frames_25k/scene0189_00/color/000000.jpg",
"scannet_frames_25k/scene0189_00/color/000200.jpg",
"scannet_frames_25k/scene0189_00/color/000400.jpg",
"scannet_frames_25k/scene0189_00/color/000600.jpg",
"scannet_frames_25k/scene0189_00/color/000800.jpg",
"scannet_frames_25k/scene0189_00/color/001000.... | This is a 3D scenario. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7][Image 8]
What color wall is the whiteboard against?
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only t... | Understand: The question asks for the color of the wall behind the whiteboard, based on the images provided. The answer should be a color description.
Perceive:
- Image 1: Shows a table and chairs, light-colored wall in the background, but no clear whiteboard.
- Image 2: Shows the same setup as Image 1, with a light-c... | red wall | gpt-4.1 | pass | true |
contrastive_caption_40ffd04ff4cb_1 | Retrieval | contrastive_caption | image_text_matching | [
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/LaionGPT4V-LIVIS_images/11773.jpg",
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/LaionGPT4V-LIVIS_images/11774.jpg",
"contrastive_caption/0.0.0/c55... | [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7]
Which image is the caption about?
a baseball player in action, taking a swing at a baseball with his bat. There are two people visible in the image: a baseball player and a catcher. The baseball player has the name "FIELDER" and the number "28" on the back... | Understand: The question asks which image matches the description of a baseball player named "FIELDER" with number "28," swinging at a baseball in front of a catcher, with stadium equipment and surroundings.
Perceive:
- Image 1: Indoor living room scene, no baseball activity.
- Image 2: Outdoor scene by a lake, no base... | 5 | gpt-4.1 | pass | true |
contrastive_caption_1d6ec3eaa736_4 | Retrieval | contrastive_caption | image_text_matching | [
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/LaionGPT4V-LIVIS_images/22951.jpg",
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/LaionGPT4V-LIVIS_images/22952.jpg",
"contrastive_caption/0.0.0/c55... | [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7][Image 8]
Which image is the caption describing?
a professional tennis match, featuring a male tennis player dressed in blue sportswear and wearing a headband. He is right-handed and preparing to hit a tennis ball with his racket on a green indoor tennis co... | Understand: The question asks for the image showing a professional tennis match with a male player in blue sportswear and a headband, right-handed, preparing to hit a ball on a green indoor court with a blue surface, surrounded by banners (including Mercedes-Benz and adidas), spectators, and flower arrangements.
Percei... | 3 | gpt-4.1 | fail | null |
jigsaw_f4ae40da06a9_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_f4ae40da06a9_0_hole.jpg",
"synthetic/jigsaw/jigsaw_f4ae40da06a9_0_p0.jpg",
"synthetic/jigsaw/jigsaw_f4ae40da06a9_0_p1.jpg",
"synthetic/jigsaw/jigsaw_f4ae40da06a9_0_p2.jpg",
"synthetic/jigsaw/jigsaw_f4ae40da06a9_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four image patches (Image 2, 3, 4, or 5) fits into the grey rectangle covering part of the scene in Image 1. The answer must be a single option letter corresponding to the correct patch.
Perceive:
- Image 1: The grey rectangle is in the central lower portion, covering part of... | B | gpt-4.1 | pass | true |
VISION_0cf82af17b70_0 | Similarity | VISION | same_variety_truefalse | [
"VISION/full/images/4338.jpg",
"VISION/full/images/4339.jpg"
] | Are these two workpieces the same genre? [Image 1][Image 2]
Choice list:
A: True
B: False
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the evidence in that image which bears on the ques... | Understand: The question asks whether the two workpieces shown in Image 1 and Image 2 belong to the same genre, meaning if they are essentially similar in material, texture, and appearance. The answer must be a choice between "True" or "False."
Perceive:
- Image 1: The workpiece displays a wood texture with visible gr... | A | gpt-4.1 | pass | true |
contrastive_caption_2f6685a938ba_0 | Retrieval | contrastive_caption | image_text_matching | [
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/sharegpt4v/coco/train2017/000000243333.jpg",
"contrastive_caption/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/sharegpt4v/coco/train2017/000000243336.jpg",
"contrastive... | Which image is the caption describing?
In the image, a zebra, adorned with its distinctive black and white stripes, is captured in mid-stride, heading towards the left side of the frame. The zebra is walking on a dirt path that is speckled with patches of grass. The path is enclosed by a chain link fence, suggesting a ... | Understand: The question asks which image matches the caption describing a zebra walking on a dirt path with grass patches, enclosed by a chain link fence, and a solitary tree in the background.
Perceive:
- Image 1: People eating at a table outdoors, no zebra or dirt path.
- Image 2: Two people at the beach with surfbo... | 5 | gpt-4.1 | pass | true |
jigsaw_18daec03d421_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_18daec03d421_0_hole.jpg",
"synthetic/jigsaw/jigsaw_18daec03d421_0_p0.jpg",
"synthetic/jigsaw/jigsaw_18daec03d421_0_p1.jpg",
"synthetic/jigsaw/jigsaw_18daec03d421_0_p2.jpg",
"synthetic/jigsaw/jigsaw_18daec03d421_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which patch (from options A-D) was cut from the region covered by the grey rectangle in Image 1. The answer must be a single option letter corresponding to the correct patch.
Perceive:
- Image 1: The grey rectangle covers a portion of the parking lot scene with cars and sky visible in the... | A | gpt-4.1 | pass | true |
imagecode_1792a1209c20_0 | Retrieval | imagecode | fine_grained_retrieval | [
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/MSR-VTT-videoTrainValVideo_video2484-shot10_0/img0.jpg",
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/MSR-VTT-videoTrainValVideo_video2484-shot10_0... | Given a detailed description, retrieve the target image among 10 minimally contrastive images
Description:
A man looks straight ahead while holding up three fingers on each hand.
A. [Image 1]
B. [Image 2]
C. [Image 3]
D. [Image 4]
E. [Image 5]
F. [Image 6]
G. [Image 7]
H. [Image 8]
I. [Image 9]
J. [Image 10]
Reason st... | Understand: The question asks for the image where a man is looking straight ahead while holding up three fingers on each hand. The answer should be the number of the image that best fits this description.
Perceive:
- Image 1: The man is looking straight ahead and appears to be holding up three fingers on each hand, wit... | 1 | gpt-4.1 | fail | null |
ScanQA_449a1f217f82_0 | Multi-view | ScanQA | multi_view_3d_qa | [
"scannet_frames_25k/scene0029_00/color/000000.jpg",
"scannet_frames_25k/scene0029_00/color/000300.jpg",
"scannet_frames_25k/scene0029_00/color/000600.jpg",
"scannet_frames_25k/scene0029_00/color/000900.jpg",
"scannet_frames_25k/scene0029_00/color/001200.jpg",
"scannet_frames_25k/scene0029_00/color/001500.... | This is a 3D scenario. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7][Image 8]
What is next to the trash can?
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the evidence ... | Understand: The question asks for the object next to the trash can in this 3D scenario, and expects a short answer.
Perceive:
- Image 1: The trash can is visible; next to it is a toilet.
- Image 2: The trash can is visible; next to it is a toilet.
- Image 3: No trash can visible.
- Image 4: No trash can visible.
- Ima... | toilet | gpt-4.1 | pass | true |
ScanQA_aa3b8a45297d_0 | Multi-view | ScanQA | multi_view_3d_qa | [
"scannet_frames_25k/scene0125_00/color/000000.jpg",
"scannet_frames_25k/scene0125_00/color/000100.jpg",
"scannet_frames_25k/scene0125_00/color/000200.jpg",
"scannet_frames_25k/scene0125_00/color/000300.jpg",
"scannet_frames_25k/scene0125_00/color/000400.jpg",
"scannet_frames_25k/scene0125_00/color/000500.... | This is a 3D scenario. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7][Image 8]
What shape is the table shaped like?
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the evi... | Understand: The question asks for the shape of the table by referring to the arrangement and visual stance across multiple images. The answer should be a brief geometric description.
Perceive:
- Image 1: Shows a long rectangular table with a straight edge, placed against the wall.
- Image 2: Shows the same table from ... | long rectangle | gpt-4.1 | fail | null |
ordering_7995b66782a9_0 | Ordering | synthetic_ordering_RecipeQA_ImageCoherence | sequence_ordering | [
"RecipeQA_ImageCoherence/full/images/14712.jpg",
"RecipeQA_ImageCoherence/full/images/14714.jpg",
"RecipeQA_ImageCoherence/full/images/14713.jpg",
"RecipeQA_ImageCoherence/full/images/14715.jpg"
] | The 4 images are the steps of a recipe, shown in a scrambled order.
Which option gives the correct order, first to last?
A. Image 3 -> Image 2 -> Image 4 -> Image 1
B. Image 2 -> Image 4 -> Image 1 -> Image 3
C. Image 1 -> Image 2 -> Image 4 -> Image 3
D. Image 1 -> Image 3 -> Image 2 -> Image 4
Reason step by step in... | Understand: The question asks for the correct chronological order of the recipe steps shown in the images. The answer must be the option letter that correctly sequences these steps from first to last.
Perceive:
- Image 1: Shows someone cutting meat or fat, with chunks and skin present—a preparation step.
- Image 2: Sh... | D | gpt-4.1 | pass | true |
jigsaw_f971b65c3ead_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_f971b65c3ead_0_hole.jpg",
"synthetic/jigsaw/jigsaw_f971b65c3ead_0_p0.jpg",
"synthetic/jigsaw/jigsaw_f971b65c3ead_0_p1.jpg",
"synthetic/jigsaw/jigsaw_f971b65c3ead_0_p2.jpg",
"synthetic/jigsaw/jigsaw_f971b65c3ead_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which patch (A, B, C, or D) was cut from the grey rectangle region in Image 1, using evidence from the provided images and matching through option-to-image mapping.
Perceive:
- Image 1: The grey rectangle covers the upper left area, overlapping part of a person's body and arm with a backg... | C | gpt-4.1 | pass | true |
ordering_58b0e47a62a1_0 | Ordering | synthetic_ordering_RecipeQA_ImageCoherence | sequence_ordering | [
"RecipeQA_ImageCoherence/full/images/47922.jpg",
"RecipeQA_ImageCoherence/full/images/47923.jpg",
"RecipeQA_ImageCoherence/full/images/47921.jpg",
"RecipeQA_ImageCoherence/full/images/47920.jpg"
] | The 4 images are the steps of a recipe, shown in a scrambled order.
Which option gives the correct order, first to last?
A. Image 2 -> Image 4 -> Image 3 -> Image 1
B. Image 1 -> Image 3 -> Image 4 -> Image 2
C. Image 4 -> Image 3 -> Image 1 -> Image 2
D. Image 1 -> Image 2 -> Image 4 -> Image 3
Reason step by step in... | Understand: The question asks for the correct chronological order of the steps shown in the four images for a recipe. The answer must be one of the given options (A, B, C, or D).
Perceive:
- Image 1: Bread slices, peanut butter, jelly, and knives are shown; no pizza ingredients or cooked food appear.
- Image 2: A cook... | C | gpt-4.1 | fail | null |
nuscenes_fc999172429d_0 | Multi-view | nuscenes | multi_view_driving_qa | [
"nuscenes/full/images/33606.jpg",
"nuscenes/full/images/33607.jpg",
"nuscenes/full/images/33608.jpg",
"nuscenes/full/images/33609.jpg",
"nuscenes/full/images/33610.jpg",
"nuscenes/full/images/33611.jpg"
] | Using six pictures taken from distinct cameras mounted on a street view car, your task is to answer queries about the presented scenario. You must choose your answer from the Choice List. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6]Question: what color are the curbs painted?
Reason step by step in this exact... | Understand: The question asks for the color of the curbs in the scenario, requiring visual identification from the provided images.
Perceive:
- Image 1: The curb is not clearly visible.
- Image 2: The curb is painted black and white.
- Image 3: The curb is painted black and white.
- Image 4: The curb is not clearly vi... | black and white | gpt-4.1 | pass | true |
ScanQA_3937314204f6_0 | Multi-view | ScanQA | multi_view_3d_qa | [
"scannet_frames_25k/scene0569_00/color/000000.jpg",
"scannet_frames_25k/scene0569_00/color/000200.jpg",
"scannet_frames_25k/scene0569_00/color/000400.jpg",
"scannet_frames_25k/scene0569_00/color/000600.jpg",
"scannet_frames_25k/scene0569_00/color/000800.jpg",
"scannet_frames_25k/scene0569_00/color/001000.... | This is a 3D scenario. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7][Image 8]
On what surface is the keyboard located?
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the... | Understand: The question asks for the surface on which the keyboard is located, requiring identification of the physical object supporting the keyboard in this room scenario.
Perceive:
- Image 1: Keyboards are placed on a flat, light-colored surface along with computers; appears to be a desk or table.
- Image 2: Shows ... | table | gpt-4.1 | pass | false |
nuscenes_78a47dec0a84_0 | Multi-view | nuscenes | multi_view_driving_qa | [
"nuscenes/full/images/8352.jpg",
"nuscenes/full/images/8353.jpg",
"nuscenes/full/images/8354.jpg",
"nuscenes/full/images/8355.jpg",
"nuscenes/full/images/8356.jpg",
"nuscenes/full/images/8357.jpg"
] | Presented with six images from various cameras on a street view car, your responsibility is to answer inquiries about the scene. You must choose your answer from the Choice List. [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6]Question: what color backpack is the guy with the blue shirt wearing?
Reason step by s... | Understand: The question asks for the color of the backpack worn by the man with the blue shirt. The answer must be a color, determined through visual evidence.
Perceive:
- Image 1: No visible people; only a truck and traffic cones.
- Image 2: A man with a blue shirt is walking across the street. He is wearing a black ... | black | gpt-4.1 | pass | true |
jigsaw_ad392b258cc7_0 | Jigsaw | synthetic_jigsaw | patch_completion | [
"synthetic/jigsaw/jigsaw_ad392b258cc7_0_hole.jpg",
"synthetic/jigsaw/jigsaw_ad392b258cc7_0_p0.jpg",
"synthetic/jigsaw/jigsaw_ad392b258cc7_0_p1.jpg",
"synthetic/jigsaw/jigsaw_ad392b258cc7_0_p2.jpg",
"synthetic/jigsaw/jigsaw_ad392b258cc7_0_p3.jpg"
] | [Image 1] has a grey rectangle covering part of the scene. One of the four patches below was cut from exactly that region; the others were cut from elsewhere in the same photo.
Which patch fills the covered region?
A. [Image 2]
B. [Image 3]
C. [Image 4]
D. [Image 5]
Reason step by step in this exact format:
Understand... | Understand: The question asks which of the four patches (given as Images 2-5) fills the grey rectangle covering part of Image 1. The answer must name the correct patch by its option letter.
Perceive:
- Image 1: The grey rectangle is positioned over a region with some red, yellow, and black colors and part of a pattern... | A | gpt-4.1 | pass | true |
VISION_807c626967ee_0 | Similarity | VISION | same_variety_truefalse | [
"VISION/full/images/278.jpg",
"VISION/full/images/279.jpg"
] | Are these two workpieces the same species? [Image 1][Image 2]
Choice list:
A: True
B: False
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the evidence in that image which bears on the qu... | Understand: The question asks whether the two workpieces shown in Image 1 and Image 2 are the same species, requiring a comparison of their visible structural characteristics to choose between True (A) and False (B).
Perceive:
- Image 1: Workpieces appear as circular objects with dark centers and lighter edges, each sh... | A | gpt-4.1 | pass | false |
ordering_11c608e9fb92_0 | Ordering | synthetic_ordering_RecipeQA_ImageCoherence | sequence_ordering | [
"RecipeQA_ImageCoherence/full/images/50866.jpg",
"RecipeQA_ImageCoherence/full/images/50864.jpg",
"RecipeQA_ImageCoherence/full/images/50867.jpg",
"RecipeQA_ImageCoherence/full/images/50865.jpg"
] | The 4 images are the steps of a recipe, shown in a scrambled order.
Which option gives the correct order, first to last?
A. Image 2 -> Image 4 -> Image 3 -> Image 1
B. Image 2 -> Image 1 -> Image 4 -> Image 3
C. Image 2 -> Image 4 -> Image 1 -> Image 3
D. Image 4 -> Image 3 -> Image 1 -> Image 2
Reason step by step in... | Understand: The question asks for the correct chronological order of the four images as steps in a recipe. The answer must be one of the given options (A, B, C, or D).
Perceive:
- Image 1: Shows completed s'mores on a plate, likely the final step of the recipe.
- Image 2: Flour is being poured into a mixer with other ... | C | gpt-4.1 | pass | true |
VISION_b8e19d31fd3f_0 | Similarity | VISION | same_variety_truefalse | [
"VISION/full/images/4604.jpg",
"VISION/full/images/4605.jpg"
] | Are these two workpieces the same species? [Image 1][Image 2]
Choice list:
A: True
B: False
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the evidence in that image which bears on the qu... | Understand: The question asks if the two workpieces shown in Image 1 and Image 2 are the same species, requiring a comparison of visible features to determine similarity.
Perceive:
- Image 1: The workpiece is mostly clear with minor spots and contamination, and a rectangular mark near the top.
- Image 2: The workpiece ... | A | gpt-4.1 | pass | false |
VISION_fe331f389dcd_0 | Similarity | VISION | same_variety_truefalse | [
"VISION/full/images/2650.jpg",
"VISION/full/images/2651.jpg"
] | Are these two workpieces the same sort? [Image 1][Image 2]
Choice list:
A: True
B: False
Reason step by step in this exact format:
Understand: state what the question asks and what form the answer must take.
Perceive: one line per image, "- Image k: ...", giving only the evidence in that image which bears on the quest... | Understand: The question asks whether the objects shown in Image 1 and Image 2 are the same sort, meaning whether they are of the same type or category. The answer must be either A (True) or B (False).
Perceive:
- Image 1: Shows a close-up of an electronic device with distinctive vent structures and a rectangular slot... | A | gpt-4.1 | pass | true |
ordering_1c24138754ec_0 | Ordering | synthetic_ordering_RecipeQA_ImageCoherence | sequence_ordering | [
"RecipeQA_ImageCoherence/full/images/38834.jpg",
"RecipeQA_ImageCoherence/full/images/38832.jpg",
"RecipeQA_ImageCoherence/full/images/38833.jpg",
"RecipeQA_ImageCoherence/full/images/38835.jpg"
] | The 4 images are the steps of a recipe, shown in a scrambled order.
Which option gives the correct order, first to last?
A. Image 2 -> Image 1 -> Image 3 -> Image 4
B. Image 2 -> Image 1 -> Image 4 -> Image 3
C. Image 4 -> Image 1 -> Image 2 -> Image 3
D. Image 2 -> Image 3 -> Image 1 -> Image 4
Reason step by step in... | Understand: The question asks for the correct chronological order of the recipe steps shown in the images, from first to last. The answer must be one of the given option letters.
Perceive:
- Image 1: Shows a finished tray of chocolate or fudge, likely a final product.
- Image 2: Shows pieces of butter in a pot, which i... | D | gpt-4.1 | fail | null |
imagecode_325154285a7e_0 | Retrieval | imagecode | fine_grained_retrieval | [
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/open-images-450_c0a17fd85b7d105d/img0.jpg",
"imagecode/0.0.0/c55fcb268ead378049e4743c77ca2db3142e12a0f7dfc42eb8267e08efa85f58/train_images/image-sets/open-images-450_c0a17fd85b7d105d/img1.jpg",
"imagecode/0... | [Image 1][Image 2][Image 3][Image 4][Image 5][Image 6][Image 7][Image 8][Image 9][Image 10]Given a detailed description, retrieve the target image among 10 minimally contrastive images
Description:
A white board with only light blue markings. Drawing of a character with a large mustache.
Reason step by step in this ex... | Understand: The question asks for the image showing a white board with only light blue markings and a drawing of a character with a large mustache. The answer must be the image number.
Perceive:
- Image 1: White board with blue drawings and text, but contains a whale, not a mustached character.
- Image 2: White board ... | 9 | gpt-4.1 | pass | true |
MIT-States_StateCoherence_18c33805269a_0 | Similarity | MIT-States_StateCoherence | state_coherence_tf | [
"MIT-States_StateCoherence/full/images/5536.jpg",
"MIT-States_StateCoherence/full/images/5537.jpg",
"MIT-States_StateCoherence/full/images/5538.jpg",
"MIT-States_StateCoherence/full/images/5539.jpg"
] | I will provide you with two sets of pictures, each of which shows an object in the opposite state. Can you tell me if the states of these two sets of pictures are the same? You must choose your answer from the Choice List. [Image 1][Image 2][Image 3][Image 4]
Choice list:
A: True
B: False
Reason step by step in this e... | Understand: The question asks whether the states depicted in two sets of pictures (images of an object in one state and an opposite state) are the same between the sets. The answer must be "True" or "False".
Perceive:
- Image 1: Shows figs growing on a tree, indicating the fig is unpicked.
- Image 2: Shows a fig place... | B | gpt-4.1 | fail | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.