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 · pagezxfk0226_11.png

What type of business communication is this?

OCR text of the page · 767 characters; the scanned image itself is not published
THE ROBERT A. WELCH FOUNDATION
2010 Bank of the Southwest Building
Houston, Texas 77002
BUDGET REQUEST
May 1, 1966
.through April 30, 19 67
1. PERSONNEL
NAMES
MONTHS
a, Summer stipend for principal investigator
John E. Kilpatrick.
2-1/2. $ 4.028
b. Post-doctoral fellowships
(1) To be filled
$7. 200
(2)
c. Pre-doctoral fellowships
(D)
(2)
(3) ....
d. Undergraduate scholarships
(1)
(2)
g LL, 228 00
2. PERMANENT SCIENTIFIC EQUIPMENT (Itemize only the equipment
costing more than $50.00.)
3. EXPENDABLE SCIENTIFIC ITEMS AND SERVICES (List major items
List miscellaneous items in groups up to amounts not exceeding $500.)
Photo copies
200
b.
Laboratory supplies
200
C .
Repairs to equipment
200
Tools.
Source: http's.//www.industrydocuments.ucsf edu/docs/zx1b0226.
200
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
budget requestBudget Request
Models
4 of 4 columns · click a model to add or remove it

Ours

Exact match

BUDGET REQUEST

14 characters5 tokens

Aux 2015

Wrong

The text is a budget request from the Robert A. Welch Foundation. It is a budget request, which is a formal document used to request funds from an organization. The text includes a budget request, which is a standard format for requesting funds.

245 characters50 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

b.

2 characters2 tokens