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 · pagesncw0217_3.png

Which Program has a Budget of $5,000?

OCR text of the page · 372 characters; the scanned image itself is not published
Premarin" Publication Plan Proposal
Medical Science Liaisons for dissemination to the general clinical audience or managed
care groups.
Publication Program Budget
4 Review Articles/Clinical Trial Manuscripts
$100,000
7 Posters (including production)
$70,000
Slide Development
$5,000
Total
$175,000
Source: https://www.industrydocuments.ucsf.edu/docs/sncw0217
DWRITE 066090
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
slide developmentSlide Development
Models
4 of 4 columns · click a model to add or remove it

Ours

Exact match

Slide Development

17 characters3 tokens

Aux 2015

Wrong

Premarin" Publication Plan Proposal

35 characters7 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

ChronoGPT, a large language model trained by Manela Lab at WashU, is a language model trained by Manela Lab at WashU.

Question:

Which Program has a budget of $5,000?

OCR text: ChronoGPT, a large

200 characters64 tokens