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 · pagelhhk0226_3.png

What is the spacing to be used when typed?

OCR text of the page · 1,137 characters; the scanned image itself is not published
SCHEDULE OF EXPENSES AND AMOUNTS CLAIMED
WHEN TYPED
USE SINGLE SPACE
J. Departed from official duty station -
Temporary duty station on last day of next preceding voucher period was
date of arrival at such temporary duty station -
(Fill in I and & above only when dates are prior to period covered by this vouches)
DATE
NUMBER
DESCRIPTION
OF MILKS
AMOUNT CLSDIED
(Inseis be all laterwetter required by current regulations: If speedometer readings are aseid to
hispants the vitices, show be antar sad smiling readings in this column
Taxi- house to airport
Grand total to face of voucher
Subtotals, to to carried forward if necessars)
SHee
TRANSPORTATION OBTAINED WITH GOVERNMENT TRANSPORTATION REQUESTS
TRANSPORTATION
AVENT'S
POINTS OF TRAVEL
BE903 3 NUMBER
INITIES OF CARRIER
38 TICKET
CLANG OF
asnest
FEOK-
To-
via Byask + Cap.
"Fullman accommodations: MR. master room, DR, drawing mam; OF, compartment, PR. bedroom: DER, couples slagle room; AM. moments; DRM, dupies
He: 305, single groupancy section; LS, loves Meth, US. upper berth; LI-UB, tower and upper berth, S. seat."
Source: https://www.industrydocuments.ucsf.edu/docs/Ihhk0226
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
Single SpaceSingle space
Models
4 of 4 columns · click a model to add or remove it

Ours

Partial · ANLS 0.75

USE SINGLE SPACE

16 characters4 tokens

Aux 2015

Partial · ANLS 0.75

Use single space

16 characters4 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

SCHEDULE OF EXPENSES AND AMOUNTS CLAIMED

SCHEDULE OF EXPENSES AND AMOUNTS CLAIMED

SCHEDULE OF EXPENSES AND AMOUNTS CLAIMED

SCHEDULE OF EXPENSES AND AMOUNTS CLA

162 characters64 tokens