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 · pagetgph0227_2.png

What is the first item in the index?

OCR text of the page · 285 characters; the scanned image itself is not published
INDEX
1 .
WATER FROM GOVERNMENT DRAIN DITCH
2 .
WATER FROM MAIN WATER TANK
3. WATER FOR WASHING VACUUM FILTERS
4.
WATER FOR WASHING WHITE SUGAR
. WATER IN BOILER FEED TANK
6. WATER IN BOILERS
7. DRINKING WATER FROM DEEP WELL
Source: https://www.industrydocuments.ucsf.edu/docs/tgph0227
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
Water from government drain ditch1. Water from government drain ditch
Models
4 of 4 columns · click a model to add or remove it

Ours

Exact match

WATER FROM GOVERNMENT DRAIN DITCH

33 characters11 tokens

Aux 2015

Wrong

1 .

3 characters3 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

1 .

3 characters2 tokens