A vision-language model read each image together with its QA and judged the item. The pass is not a filter that only removes rows — it rewrites answers it finds wrong but salvageable, drops what it cannot salvage, and adds QA where the image supports more than the source provided. Each row carries the judge's own record in clean_meta, including the cleaner identity, the policy it applied, and its per-item scores for legibility, richness and coverage.

A large share of the QA here was written by the cleaning pass, not by the original dataset. Across the pool that share runs from roughly half to over 80% of a family's turns, and it is reported in the table above. Those rows have an empty org_answer because no pre-clean original exists. Treat them as model-generated supervision: they were judged against the image, but they are not human annotation, and model-written QA is where formatting defects are most likely.

The effect on the answers that were carried over is substantive rather than cosmetic. In this pool the pass corrects values, not just wording — a curve's critical point restated from 4.00 to 2.00, a computed ratio from 1 to Approximately 1.33 — and for some families it removed the content entirely, which is why those families are absent here rather than published empty.

`answer` vs `org_answer`

answer — the cleaned answer. Train on this.
org_answer — the pre-cleaning answer from the same (image, question) in the uncleaned pool. It is empty for QA the cleaning pass added, which have no original.

Where the two differ, the difference is the correction. Keeping both makes every edit auditable instead of asking you to take the cleaning on trust.

Schema

image — HF Image(), renders directly in the dataset viewer
image_sha256 — content hash of the image
n_qa — number of turns attached to this image
qa — list of {question, answer, org_answer}, in source order
family / clean_meta — provenance and the judge's per-row record

A row with more than one entry in qa is a conversation over a single image, not a set of independent questions — the turns depend on each other and the image belongs to all of them. Keep them together and in order when training.

Parquet row groups are written at ~32 MB so the dataset viewer pages without stalling.

What is NOT claimed

No rejection sampling was run against this family: there is no accept/reject verdict per row, and answer is the cleaned reference rather than a model generation that a string verifier confirmed. For that, use the *_RS_think / *_rejected datasets in elliot-mllm.

Examples

10 rows taken straight from this dataset. Images are the original files as stored here - no downscaling, no recompression - and every turn is shown in full.

Example 1 — 3 turns on one image

#	question	answer
1	Please write out the expression of the formula in the image using LaTeX format.	V _ { t } ^ { 2 } - V _ { 0 } ^ { 2 } = 2 a x
2	Is the text in the image typed or handwritten?	Handwritten
3	What is the background of the image?	A plain, light-colored surface, likely paper.

Example 2 — 2 turns on one image

#	question	answer (cleaned)	org_answer (pre-judge)
1	Please write out the expression of the formula in the image using LaTeX format.	20% = \frac{1}{5}	2 0 % = \frac { 1 } { 5 }
2	What is the writing style of the mathematical expression in the image?	The expression is handwritten in black ink.	(unchanged)

Example 3 — 2 turns on one image

#	question	answer
1	Please write out the expression of the formula in the image using LaTeX format.	y _ { 2 } = - \frac { 8 } { 3 }
2	Is the equation in the image typed or handwritten?	The equation is handwritten.

Example 4 — 2 turns on one image

#	question	answer
1	What is the full mathematical expression shown in the image, including the introductory text?	解：\vec{DE} = 2
2	What is the value assigned to the vector DE in the equation?	2

Example 5 — 2 turns on one image

#	question	answer (cleaned)	org_answer (pre-judge)
1	Please write out the expression of the formula in the image using LaTeX format.	BQ = AB \cdot \tan A = 20\sqrt{3} cm	B Q = A B \cdot \tan \angle A = 2 0 \sqrt { 3 } c m
2	Describe the visual style of the text and the background.	The text is handwritten in black ink on a white background featuring a horizontal dotted line at the bottom.	(unchanged)

Example 6 — 3 turns on one image

#	question	answer
1	What unit of measurement is indicated by the symbol at the very end of the equation?	Ohms ($\Omega$).
2	What is visible in the top right corner of the image?	A red mark, likely a checkmark or correction.
3	On what type of surface is the formula written?	Lined paper.

Example 7 — 2 turns on one image

#	question	answer (cleaned)	org_answer (pre-judge)
1	Please write out the expression of the formula in the image using LaTeX format.	( 2 ^ { 2 } + 3 ^ { 2 } + \cdots + 49 ^ { 2 } )	( 2 ^ { 2 } + 3 ^ { 2 } + \cdots + 4 9 ^ { 2 } )
2	What is the visual style and medium of the text in the image?	The text is handwritten in dark ink on a white background.	(unchanged)

Example 8 — 2 turns on one image

#	question	answer (cleaned)	org_answer (pre-judge)
1	Please write out the expression of the formula in the image using LaTeX format.	3a = 4a + 8 - 3	3 a = 4 a + 8 - 3
2	Is the text in the image printed or handwritten?	The text is handwritten.	(unchanged)

Example 9 — 3 turns on one image

#	question	answer
1	Please write out the expression of the formula in the image using LaTeX format.	\therefore c = 0 d = 3
2	What color is the ink used for the handwritten text?	Blue
3	Which character in the expression is written with a noticeably thicker or bolder stroke?	The letter 'd'

Example 10 — 2 turns on one image

#	question	answer (cleaned)	org_answer (pre-judge)
1	Please write out the expression of the formula in the image using LaTeX format.	= -25 + 8 \frac{1}{5}	= - 2 5 + 8 \frac { 1 } { 5 }
2	What is the visual style and medium of the text shown in the image?	The text is handwritten in black ink on a white background.	(unchanged)

Downloads last month: 28

Total file size:

3.47 GB

You need to agree to share your contact information to access this dataset

hme100k_train_cleaned

How this was cleaned

answer vs org_answer