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 · pagejhnm0081_3.png

Whose daughter should you never date?

OCR text of the page · 531 characters; the scanned image itself is not published
SMOOTH
Smooth
characte
Foolproof
dating advice
imp
a
SMOOTH
1. Never date a woman named Dave.
1. Run into the
2. Never date a woman who has
to the shore, as
The more she :
2 older brothers
3. Never date a woman whose
2. If you're over
father is called "Captain."
a towel with ve
4. Always break the ice
3. If you're und
by offering her a Camel.
a towel with he
4. Always have
eady when ;;
party begins.
51359 4334
SURGEON GENERAL'S WARNING: Cigarette
Smoke Contains Carbon Monoxide.
LOOK FOR MORE
SMOOTH MOVES
COMING YOUR
WAY SOON.
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
"captain"Captainfather is called "captain"Never date a woman whose father is called "Captain".
Models
4 of 4 columns · click a model to add or remove it

Ours

Partial · ANLS 0.77

a woman whose father is called "Captain"

40 characters10 tokens

Aux 2015

Wrong

SMOOTH

6 characters4 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

Smooth Smooth Foolproof dating advice imp a SMOOTH

51 characters20 tokens