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 · pagerrnc0227_32.png

What is the percentage of meat required by the airforce at location Peshawar?

OCR text of the page · 899 characters; the scanned image itself is not published
TABLE 4
Pakistan:
Kitchen Preparation Refuse (Percent)
Cauli-
Onion/
Spin-
Meat
Fish Potato
Turnip Carrot Peas Cabbage flower
ach
Fruit
Location
%
%
%
%
%
%
%
%
%
%
%
I .
Army
Lahore Med.-
22.0
26.3
24.8
Lahore Inf. 3/
33 . 6
26.0
34.4
28.6
44.1
18.8
Quetta 3/ .
21.0
19.7
16.8-
18.0
15. 4
19.1
22.2
18. 1
Bahawalpur 3/
18.6
39. 6
8. 8
65.2
16.4
1
33 . 4
35.0
4.0
Muzaffarabad 3/27.6
34.4
68.4
31.7
50.5
Comilla, E.
Bengal 4/
Estimated
Army Average -
22.6
22.3
35.6
27.9 19.9
53.3
19.1 18.1
Lahore Army Boys 23.1 13.4 5.9
35.1
24 .7
61 . 4
19.8
28.8
30.2
II.
PNS Thelum
19.9
18.9
21.0
25.7
41.6
10. 4
-
11.6
13 . 3
III. Air Force
Peshawar No.2 21.9 37.6 27.1
34.0
19.5
20.4
17. 5
42.7
1/
Larger per cent refuse with winter loose-type onion.
2/
Varies from 4 per cent for guava to 42.7 per cent for tamarind.
3/
Wheat-eaters
4/
Rice-eaters
Source: https://www.industrydocuments.ucsf.edu/docs/rrnc0227
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
21.9
Models
4 of 4 columns · click a model to add or remove it

Ours

Exact match

21.9

4 characters5 tokens

Aux 2015

Wrong

We are given a table with percentages of refuse (meat) required by the air force at various locations. We need to calculate the percentage of meat required by the air force at location Peshawar.

Let's analyze the table:

  • Army:
    • Lahore: 22.0
250 characters64 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

Percentage of meat required by the airforce at location Peshawar.

ChronoGPT: Kitchen Preparation Refuse (Percentage) Cauli- Onion/ Spin- Meat Fish Potato Turnip Carrot Peas Cabbage Flower Fruit

194 characters64 tokens