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 · pagehnjh0227_3.png

What is enclosed with this letter ?

OCR text of the page · 565 characters; the scanned image itself is not published
Dec. 15, 1970
National Corporation for Housing Partnerships
1 133 Fifteenth Street, Northwest, Washington, D. C. 20005 Tel: (202) 466-8200
Near Tune.
Carter L. Burgess
Chairman
I am enclosing the latest information
George W. DeFranceaux
President
we have on Materials Systems . I shall let
you know when I get any additional information .
I trust you will be flashing down
The slopes over the holidays . I am sorry I
Shall not be able to join you , but I shall be
out there later this mister .
Best.
hang
Source: https://www.industrydocuments.ucsf.edu/docs/hnjh0227
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
The latest information we have on materials systems.
Models
4 of 4 columns · click a model to add or remove it

Ours

Partial · ANLS 0.98

the latest information we have on Materials Systems

51 characters9 tokens

Aux 2015

Wrong

National Corporation for Housing Partnerships

133 Fifteenth Street, Northwest, Washington, D. C. 20005

Tel: (202) 466-8200

Near Tune.

Carter L. Burgess

Chairman

I am enclosing the latest

192 characters64 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

I am enclosing the latest information . I shall let you know when I get any additional information . I shall let you know when I get any additional information . I shall let you know when I get any additional information . I shall let you know when I get any additional information . I shall let you know when I get

315 characters64 tokens