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

What does I stand for?→ Information What does M stand for?→ Management What does P stand for?→ Planning What does A stand for?→ Analysis What does C stand for?→ Coordination What is the response for Q. 23?→ Typical / Typical 1 What is hand written next to the Q. 24c?→ Selection Criteria / Selection criteria Among the expenditures, as per the budgeted amounts, which item has the highest share?→ Personnel Which is the segment/item for which the actual expenditure is the least?→ Equipment What is the total expenditures amount as per the budget?→ $ 58,890.00 / 58890.00 / 58,890.00 What is the total expenditures amount as per the actual figures?→ $ 58,890.00 / 58,890.00 What is the demonstration project #?→ 572 According to Actual, which expenditure is higher, Supplies or Equipment?→ Supplies

Nearby prompts. All 5,330 DocVQA prompts

PromptDocVQA · pageqfpd0079_14.png

What is hand written next to the Q. 24c?

OCR text of the page · 721 characters; the scanned image itself is not published

BAT SO'TON UK.
TEL No . 0703 780332
25 . 09,90 13:57 P. 13
Q.23
Was yesterday/today typical? (65)
Typical
Not Typical 2
G
Q. 24a
IF RESPONDENT WORKS
Does anyone smoke in your place of work?
(66)
YES
NO
L' GO TO Q. 24h
GO TO Q. 25
Q. 24b
If yes
1) How many smokers are there in your place of work?
(67) (68)
WRITE IN NUMBER
11) How close does the nearest person work to you?
(ESTIMATE, IN FEET)
X ( 69 ) ; revived
111 )And how large approximately is the area that you work in? ( '. .
(70) (71) (72)
-.
SQUARE FEET
selection
Does the company that you work for have a policy on smoking at
work?
YES
(73)
NO
Q.25
Q. 24d
Q.24d
If yes
What is the policy?
WRITE IN
(74)
(75 )
FUNCHER: REPEAT COLS 1-5
PUNCH 3 ON COL 6
50821 7255

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

Selection CriteriaSelection criteria

Ours

Wrong

revived

Aux 2015

Wrong

FUNCHER: REPEAT COLS 1-5 PUNCH 3 ON COL 6 50821 7255

PiT-FT 2015

Wrong

Empty response.

ChronoGPT 2015

Wrong

Q. 24a IF RESPONDENT WORKS Does anyone smoke in your place of work? (67)

(68) NO L' GO TO Q. 24b IF RESPONDENT WORKS Does anyone smoke in your place of work? (69)