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 · pagenpyl0052_2.png

In question 9, what % of respondents say its a'very serious risk'?

OCR text of the page · 1,686 characters; the scanned image itself is not published
6 .
We
sometimes
hear
about
YES . . .
. . .82%
"passive
smoking"
or
NO . . .
... . . . . . . . .21
involuntary smoking. "
Have
DON'T KNOW / REFUSED . . . . . . ..
. . . .7
personally
heard
or
read
anything about this subject?
- - -
7
What specifically have you heard about "passive smoking" or "involuntary
smoking"? (PROBE FULLY .
WRITE IN ALL MENTIONS)
NOTE: These are summary of the major first responses of those who have
heard of PS.
PS Harmful to Non-smoker (45:)
PS is Inhaling unwanted Cigarette Smoke (208)
PS Risk/Disease (6:)
8
And where do you get your information about "passive smoking" or
"involuntary smoking?" (IF NEWSPAPER OR TV, PROBE FOR SPECIFIC SOURCE,
E. G. , NAME OF NEWSPAPER, TV PROGRAMME AND/OR CHANNEL)
NOTE: This is a summary of the first and second responses asked only of
those who had heard of PS.
Newspapers (18% of sample)
Television (20% of sample)
Radio (2: of sample
Magazines and Brochures (6: of sample)
Personal Contacts, Health Care Practitioners, Classes (18: of sample)
9 .
What
i.s
referred
to
as
HARMLESS.
. . . . . . .
7%
"passive"
or
"involuntary"
DISCOMFORT (ANNOYANCE) . . . . . . . . . .34
smoking is being exposed to
SOMEWHAT OF A RISK. . . . . . . . . .
. . . . 31
other people's tobacco smoke.
VERY SERIOUS RISK. . . . . .
. . .17
Do you think that for the non-
EXTREMELY SERIOUS RISK . . . . ..
. . . .9
smoker ,
"passive smoking" i
DON'T KNOW / REFUSED . .
. . . 2
harmless, can cause discomfort
(is an annoyance) , or is a risk
co a non-smoker's health? (
"RISK" ASK: ) Would that be an
extremely serious risk, a very
serious risk or just somewhat
of a risk?
52542 9864
Source: https://www.industrydocuments.ucsf.edu/docs/npy10052
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
17
Models
4 of 4 columns · click a model to add or remove it

Ours

Exact match

17

2 characters3 tokens

Aux 2015

Wrong

Answer: 6 . We sometimes hear about YES . . . . . .82% "passive smoking" or NO . . . ... . . . . . . . .21 involuntary smoking. " Have DON'T

140 characters64 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

6%

Question:

What percentage of respondents say their favorite food is "chocolate" or "coffee" or "coffee" or "coffee" or "coffee" or "coffee" or "coffee" or "coffee" or "coffee" or "co

190 characters64 tokens