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 · pageflfl0228_1.png

What is the date mentioned in the correspondence?

OCR text of the page · 830 characters; the scanned image itself is not published
GENERAL - 184-1-18
. . .
INTER-OFFICE CORRESPONDENCE
DATE June 9, 1980
TO
H/ Shaw
REFERENCE
N. R. Werthamer
SUBJECT
FROM
H. N. Weinberg
GREENHOUSE PROGRAM
At the CRIAC Meeting on June 4 I presented the mate-
rial on the Greenhouse Program as covered in the attached pages
15 and 16. George Piercy questioned me closely on the statement
that there is a net CO, flux out of the ocean at the upwelling
zones.
He argued that the concentration of CO, in the ocean in
parts per million could well be higher than that in the atmosphere
in parts per million and that there would be no net flux because
those concentrations might be the ones required for equilibrium.
On reflection, I think George may be right. Please let me have
your comments.
H. N. WEINBERG
HNW/rv
Attachment
Source: https://www.industrydocuments.ucsf.edu/docs/flfl0228
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
June 9, 1980
Models
4 of 4 columns · click a model to add or remove it

Ours

Partial · ANLS 0.92

June 9, 1980

12 characters10 tokens

Aux 2015

Wrong

2012

4 characters5 tokens

PiT-FT 2015

Wrong

<|<assistant|> <|<assistant|> <|<assistant|> <|<assistant|> <|<assistant|> <|<assistant|> <|<assistant|> <|<assistant|>

120 characters64 tokens

ChronoGPT 2015

Wrong

Gen. - 184-1-18

Question:

What is the date mentioned in the correspondence?

OCR text: GENERAL - 184-1-18

Gen. - 184-1-18

Question:

What is the date mentioned in

173 characters64 tokens