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 · pagezsdg0227_1.png

Who is it from?

OCR text of the page · 154 characters; the scanned image itself is not published
From : Dr. Hugo E. Gallo-Torres, X.D.
Pasaje San Luis 183-4
San Miguel, Lima
Peru
Sud America
Source: https://www.industrydocuments.ucsf.edu/docs/zsdg0227
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
DR. HUGO E. GALLO-TORRESDr. Hugo E. Gallo-Torres
Models
4 of 4 columns · click a model to add or remove it

Ours

Partial · ANLS 0.80

Dr. Hugo E. Gallo-Torres, X.D.

30 characters15 tokens

Aux 2015

Exact match

Dr. Hugo E. Gallo-Torres

24 characters11 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters1 tokens

ChronoGPT 2015

Partial · ANLS 0.80

Dr. Hugo E. Gallo-Torres, X.D.

30 characters15 tokens