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 · pagegfhd0082_13.png

Which card is handed to the respondent second?

OCR text of the page · 2,218 characters; the scanned image itself is not published
( HAND RESPONDENT .CARD "E" )
36.
As a brand of cigarettes for your own use, what number from 10 to 100 best describes
this cigarette you just read about?
Outstanding . . . . . . . . . 100
1 1
73-0
Excellent . . . . . . . . . . . 9
90
Very Good. . . . . . . . . ..
Quite Good. . . . . . . . ..
Fairly Good. . . . . . . . .
X
Indifferent. . . . . . . . . 50
Not Very Good . . . . . . . 40
80-1
Not Good At All. .. .. 30
poor. . . . . . . . . . . . . . ..
20
Terrible. . . . . . ..
. . . . 10
(TAKE BACK CARD "E" )
(HAND RESPONDENT CARD "A")
CARD 2
37. As before, on this sheet are pairs of phrases which might be used to describe the cigarette
you just read about. As you can see, it has one phrase on each side of the page with a
scale from 1 to 7 drawn in between each pair of phrases. Using this scale, please tell
me the position which best describes what you would expect this cigarette to be like.
(MAKE SURE RESPONDENT UNDERSTANDS BEFORE CONTINUING WITH EACH PAIR OF PHRASES. )
Not strong tasting at all. . . . ...
3
5 6 7....... .Extremely strong tasting
No strong aftertaste....
7. .... . ..Extremely strong aftertaste
Not harsh at all. . . . . . . . . . . . . . . . .
1 2
5
6
5
6
7. . . . . . . .Extremely harsh
No fresh aftertaste. . . . . . . . . . . ...
1
2 3
5
6
7. . . . .. . . Extremely fresh aftertaste
Natural taste. . .
. . . . . . .
1 2
4 5
6
.;. .. .Artificial taste
10
No long lasting aftertaste. . . . . . .
1 2 3
5 6
7. . ..... . Extremely long lasting
aftertaste
11
Not smooth at all . . . . . . . . . . . . . . . .
1
23
5
7. . . . . . . . Extremely smooth
12
No drying aftertaste. . . . . . . . . . . . .
2
6
7. . . . . . . . Extremely drying aftertaste
13
No tobacco taste at all. . ....
2
6.
7. . . . . . . .Extreme amount of tobacco
taste
No bitter taste. .
. . . . .. 1 2 3
(4' 5
6 7........Extreme amount of bitter
taste
15
Not mild at all . .. .
. . . . . . 1
2
5 6
.. . .. .Extremely mild
16
Easy to get smoke through the
filter. . . . . . . . .. .. . . . . . . . . . . . .
(2 ) 2 3 4 5 6
7.......Hard to get smoke through
the filter
-17
(TAKE BACK CARD "A" )
GO TO Q. 38 - SHOW SAMPLE CIGARETTE AND COMPLETE
.. ..
CALLBACK INFORMATION ON THE FOLLOWING PAGE.
. . .. ..
50557 9770
- 13 -
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
a"A""a"Card "A"A
Models
4 of 4 columns · click a model to add or remove it

Ours

Partial · ANLS 0.80

"Card "A""

10 characters6 tokens

Aux 2015

Wrong

The respondent is handed a card with the phrase:

"NOT STRONG TASTING AT ALL. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

163 characters64 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

(HAND RESPONDENT .CARD "E" )

CARD 2 37

CARD 1 73-0

CARD 1 73-0

CARD 1 73-0

CARD 1 73-0

CARD 1 73-

103 characters64 tokens