The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed
Error code:   DatasetGenerationError
Exception:    CastError
Message:      Couldn't cast
input: string
output: string
gts: list<item: string>
  child 0, item: string
score: double
step: int64
sample_id: string
reward: double
em_score_mcq: int64
llm_score: double
format_score: double
to
{'input': Value('string'), 'output': Value('string'), 'gts': List(Value('string')), 'score': Value('float64'), 'step': Value('int64'), 'em_score_mcq': Value('int64'), 'llm_score': Value('float64'), 'format_score': Value('float64'), 'sample_id': Value('string')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1816, in _prepare_split_single
                  for key, table in generator:
                                    ^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 613, in wrapped
                  for item in generator(*args, **kwargs):
                              ~~~~~~~~~^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2369, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2297, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              input: string
              output: string
              gts: list<item: string>
                child 0, item: string
              score: double
              step: int64
              sample_id: string
              reward: double
              em_score_mcq: int64
              llm_score: double
              format_score: double
              to
              {'input': Value('string'), 'output': Value('string'), 'gts': List(Value('string')), 'score': Value('float64'), 'step': Value('int64'), 'em_score_mcq': Value('int64'), 'llm_score': Value('float64'), 'format_score': Value('float64'), 'sample_id': Value('string')}
              because column names don't match
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1369, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1683, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1869, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

input
string
output
string
gts
list
score
float64
step
int64
em_score_mcq
int64
llm_score
float64
format_score
float64
sample_id
string
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> I need to find information about the co-founder and CEO of Canva, specifically her academic majors during her time at The University of Western Australia. Since this requires external knowledge, I will use a text search tool to look for relevant details.</thinking> <tool_call> {"name": "text_search_tool", "a...
[ "Communications, psychology and commerce" ]
1.5
1
null
1
0.5
fvqa_train-1788
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> The person in the image is Anne Hathaway, and she is known for portraying a character in the film "Verity." However, to confirm the name of the character, I need to search for more specific information about her role in the movie. I'll perform a web search to find details about her character in "Verity." </t...
[ "Verity Crawford" ]
1.5
1
null
1
0.5
fvqa_train-1766
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> The person in the image is Anne Hathaway, a well-known actress. Since the question is about the character she portrays in the film "Verity," I need to determine which character she played in that movie. I will search for information about Anne Hathaway's role in "Verity." <tool_call> {"name": "text_search_...
[ "Verity Crawford" ]
0.5
1
null
0
0.5
fvqa_train-1766
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> I need to identify who is wearing a shirt in the image. The main focus should be on the people in the foreground, as they are most visible and likely to be the subject of the question. I will zoom in on the group of people to get a clearer view of their clothing. <tool_call> {"name": "image_zoom_in_tool", "a...
[ "The woman is wearing a shirt." ]
1.5
1
null
1
0.5
deepeyes_4k_train_non_mcq-39
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> I need to determine what is covering the hills in the image. To do this, I should zoom in on the hills to get a closer look at their surface features. </thinking> <tool_call> {"name": "image_zoom_in_tool", "arguments": {"bbox_2d": [0, 562, 533, 637], "label": "the hills in the background", "img_idx": 0}} </t...
[ "The hills are covered in trees." ]
0.5
1
null
0
0.5
deepeyes_4k_train_non_mcq-193
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> I need to examine the image to determine if the tray is only partially seen to the left or right of the logo. First, I'll locate the logo and then check its surrounding area to see how the tray appears relative to it. <tool_call> {"name": "image_zoom_in_tool", "arguments": {"bbox_2d": [18, 114, 260, 473], "l...
[ "The tray only partially seen is to the right of the logo." ]
1.5
1
null
1
0.5
deepeyes_4k_train_non_mcq-403
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> The logo features a neon star with the word "BIG" inside it. This suggests the name of the company might be "BIG." To confirm, I should examine the details more closely, especially any text or symbols around the star. </thinking> <tool_call> {"name": "image_zoom_in_tool", "arguments": {"bbox_2d": [122, 45,...
[ "Big Star" ]
0.5
1
null
0
0.5
fvqa_train-365
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> To determine the architect of this building, I need to analyze the image for any identifiable details such as inscriptions, logos, or architectural styles that might point to a specific architect. However, since the building is ornate and historic, its architect may be mentioned in historical context or on p...
[ "Peter Parler" ]
0.5
1
null
0
0.5
fvqa_train-1020
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> The image depicts a 3D stacked memory structure, where multiple layers of memory cells are stacked vertically. This is indicative of a specific type of advanced memory technology. To identify the exact name, I need to analyze the structure more closely or search for related information. I will use the image_...
[ "3D XPoint" ]
1.5
1
null
1
0.5
fvqa_train-447
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> I need to identify the street address of this building. Since there is no visible signage or text in the image, I will zoom in on the building to look for any signs, plaques, or nearby street markers that might indicate the address.</thinking> <tool_call> {"name": "image_zoom_in_tool", "arguments": {"bbox_2d...
[ "piazza Bocca della Verità" ]
0.5
1
null
0
0.5
fvqa_train-1354
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> To determine the location of this lake, I need to identify key features or landmarks in the image. The presence of a hillside with buildings, dense trees, and a distinctive green bridge near the water suggests it might be a known scenic spot. I should start by zooming in on the bridge and surrounding area to...
[ "Banjosa" ]
0.5
1
null
0
0.5
fvqa_train-1017
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> To identify the person in the image, I should first examine the facial features and attire for any distinctive characteristics that might help recognize them. Zooming in on the face will give me a clearer view to analyze details such as hairstyle, facial structure, and clothing style.</thinking> <tool_call> ...
[ "Roscoe Arbuckle" ]
1.5
1
null
1
0.5
fvqa_train-210
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> To determine the year this building was first mentioned in writing, I need to gather historical information about it. Since the image shows a castle or fortified structure, I should perform a web search to find any historical records or mentions related to this specific building. </thinking> <tool_call> {"...
[ "1285" ]
0.5
1
null
0
0.5
fvqa_train-984
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> To determine who designed this building, I need to identify the specific structure in the image and find information about its architect. The most prominent building in the image appears to be the church with the twin spires in the background. I should zoom in on that area to better examine it and look for a...
[ "Helgo Zettervall" ]
0.5
1
null
0
0.5
fvqa_train-1491
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> To determine where this charming street is located, I should look for distinguishing architectural features or landmarks in the image. Zooming in on specific buildings or elements might reveal clues such as signage, unique designs, or surrounding context that can help identify the location.</thinking> <tool_...
[ "Mackinac Island" ]
1.5
1
null
1
0.5
fvqa_train-457
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> I need to identify the athlete wearing the Team GB jacket with a silver medal. The image shows a person wearing a Team GB jacket and a medal, but their identity is not immediately clear. To determine who this athlete is, I should zoom in on the medal or the jacket to read any identifying text or logos that m...
[ "Kye Whyte" ]
0.5
1
null
0
0.5
fvqa_train-656
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> The image shows a coastal scene with stone buildings featuring moss-covered roofs, and a hill in the background. This architectural style and setting suggest it might be located in a region with a significant history of maritime trade or tourism. However, without distinctive landmarks or signs, I need to use...
[ "Outer Hebrides" ]
0.5
1
null
0
0.5
fvqa_train-1484
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> To determine the architect of this building, I need to identify key architectural features and possibly look for any signage or inscriptions that might mention the architect. Zooming in on the building's facade or nearby signs could help. I will zoom in on the building to examine its details more closely.</t...
[ "Ad van der Steur" ]
1.5
1
null
1
0.5
fvqa_train-1450
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> The person in the image is wearing traditional attire that appears ceremonial, including a golden robe with intricate embroidery and a decorative hat. To identify their title or role, I should focus on their attire and context. Zooming in on the person's face and attire might reveal more clues, such as insig...
[ "chogyal" ]
0.5
1
null
0
0.5
fvqa_train-915
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> To determine the architect of this building, I need to analyze the architectural features and any visible inscriptions or markers that might identify the designer. The image shows a prominent tower with Gothic-style elements, which may help in identifying the architect through historical knowledge or labels ...
[ "Jan van Haelst" ]
1.5
1
null
1
0.5
fvqa_train-1387
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> The image shows a mountain in the background with a building or structure on top. To determine what the mountain is named after, I should zoom in on the mountain to get a closer look at any visible names or inscriptions. I will use the image_zoom_in_tool to focus on the mountain.</thinking> <tool_call> {"n...
[ "Montserrat" ]
0.5
1
null
0
0.5
fvqa_train-1668
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> To determine which diocese this building belongs to, I need to identify the specific church or cathedral in the image. The architectural style and location can provide clues. I will start by zooming in on a key feature of the building, such as the main entrance or the top of the structure, to find any inscri...
[ "Roman Catholic Archdiocese of Lucca" ]
1.5
1
null
1
0.5
fvqa_train-89
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> I need to analyze the top left graph to determine where the density of points is the highest along the u2 axis. The graph shows a cluster of red dots, and I should identify the range on the y-axis (u2) where these dots are most concentrated. Zooming in on this region will help me get a clearer view of the po...
[ "C" ]
0.5
1
0
null
0.5
deepeyes_4k_train_mcq-209
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> I need to find the median RMS noise value at which the largest area difference between "All fields" (black line) and "RA-0 fields" (orange line) occurs. This requires comparing the vertical distance between the two lines at various points along the x-axis (Median RMS Noise). The largest vertical gap will ind...
[ "C" ]
1.5
1
1
null
0.5
deepeyes_4k_train_mcq-566
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> To determine which component of Fisher Information shows the most variability, I need to analyze the fluctuations or spread of each line on the graph. The “Total Infor” line appears to have the steepest initial rise and then stabilizes, while “First Response Infor” and “Revision Infor” show different pattern...
[ "A" ]
0.5
1
0
null
0.5
deepeyes_4k_train_mcq-581
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> The question is asking for the background color of "Original," which appears to be a text label on a framed picture or object in the image. I need to locate that specific item and examine its background color closely. To do this, I will zoom in on the framed picture with the text "Original" to get a clearer ...
[ "white" ]
0.5
1
null
0
0.5
visual_probe_train-870
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> I need to identify the sculpture on top of the building. Since it's located at the peak of the structure, zooming in will help me get a clearer view and accurately determine what animal it is. I'll start by zooming in on that specific area.</thinking> <tool_call> {"name": "image_zoom_in_tool", "arguments": {...
[ "phoenix" ]
1.5
1
null
1
0.5
visual_probe_train-50
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> I need to locate the word "AMORE" in the image and identify the word immediately before it. The text appears to be part of a newspaper headline or article, so I will first zoom in on the relevant section to read it clearly. </thinking> <tool_call> {"name": "image_zoom_in_tool", "arguments": {"bbox_2d": [784,...
[ "IN" ]
0.5
1
null
0
0.5
visual_probe_train-838
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> I need to examine the decorative wall behind the bed, specifically focusing on the first plank in the upper right corner. To determine if its pattern is horizontal or vertical, I must zoom in on that area for a closer inspection. </thinking> <tool_call> {"name": "image_zoom_in_tool", "arguments": {"bbox_2d":...
[ "horizontal" ]
0.5
1
null
0
0.5
visual_probe_train-534
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> I need to locate the cabinet in the middle-right of the image and identify the text on the spine of the topmost book. The text might be small, so zooming in will help read it accurately. I will use the image zoom-in tool on the cabinet area to get a closer look at the books. <tool_call> {"name": "image_zoom_...
[ "WHEN ART MEETS DESIGN" ]
1.5
1
null
1
0.5
visual_probe_train-373
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> I need to locate the second row from the top and the leftmost book in that row. Then, I must identify the number written on its spine. To ensure accuracy, I will zoom in on that specific area. <tool_call> {"name": "image_zoom_in_tool", "arguments": {"bbox_2d": [0, 250, 492, 370], "label": "second row from th...
[ "100" ]
0.5
1
null
0
0.5
visual_probe_train-209
<|im_start|>system #Role You are a step-by-step reasoning assistant. Given a question, your task is to solve the problem **one substep at a time**. ## Guiding Principles At each turn, you must **either**: 1. Issue **one specific tool** enclosed in <tool_call> </tool_call> tags, 2. Or provide the **final answe...
<thinking> I need to count the white tea cups on the coffee table. The coffee table is in the center of the room and has a tray with several items on it. Zooming in on the coffee table will help me clearly see the items, especially the tea cups. </thinking> <tool_call> {"name": "image_zoom_in_tool", "arguments": {"bbox...
[ "Four" ]
1.5
1
null
1
0.5
visual_probe_train-440
End of preview.