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 · pagesphv0228_21.png

In which calendar months were the number of workers the highest?

OCR text of the page · 194 characters; the scanned image itself is not published
25
22
14 WITH 7 TO 25 MO.
8 WITH 26 TO 50 MO.
NO. OF WORKERS
20
1975 STUDY
15
10
9
10
7
7
5
O
0-50 51-100 101-150 151-200 201-250 (261 AND 310)
( CALENDAR MONTHS OF DITHANE EMPLOYMENT )
FIGURE 7
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
0-50
Models
4 of 4 columns · click a model to add or remove it

Ours

Wrong

261-310

7 characters8 tokens

Aux 2015

Wrong

25

22

14

7

7

5

O

0-50 51-100 101-150 151-200 201-250 (261 AND 310)

(

76 characters64 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

25 22 14 8 10 7 5 O 0-50 51-100 101-150 151-200 201-250 (261 AND 310)

69 characters36 tokens