Output Explorer

Every prompt in the paper, and what each model wrote back.

Answer a question about a scanned page from its OCR text. Scored by exact match and ANLS against the accepted answers.

13 of 5,330 prompts

Nearby prompts. All 5,330 DocVQA prompts

PromptDocVQA · pagexnyc0227_2.png

What type of report is this?

OCR text of the page · 161 characters; the scanned image itself is not published
Shanle
PRELIMINARY
NORTHEAST
BRAZIL
NUTRITION SURVEY
March - May 1963
A Report by the Interdepartmental Committee on Nutrition for National Defense
February 1964
System prompt · identical for every setup
Answer the question using only the OCR text from a single document page. Return only the answer, with no explanation. Preserve the answer wording from the OCR text when possible.
Expected answer
Preliminary
Models
4 of 4 columns · click a model to add or remove it

Ours

Wrong

PRELIMINARY NORTHEAST BRAZIL NUTRITION SURVEY

45 characters19 tokens

Aux 2015

Exact match

PRELIMINARY

11 characters5 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters1 tokens

ChronoGPT 2015

Wrong

OCR text: Shanle PRELIMINARY NORTHEAST BRAZIL NUTRITION SURVEY March - May 1963 A Report by the Interdepartmental Committee on Nutrition for National Defense February 1964

171 characters56 tokens