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 · pagepyyc0227_52.png

What is the minimum age group shown in the ‘plots’?

OCR text of the page · 820 characters; the scanned image itself is not published
507
TOTAL GROUP
BOYS
MALNOURISHED
Marasmus,
48-
Prekwashiorkor,
<washiorkor)
46-
( 21
44-
52 -
. (2)
- BOYS TOTAL
50 -
- GIRLS
GROUP
42-
cm
MALNOURISHED BOYS
40-
48 -
38-
46 -
36-
44-
MALNOURISHED GIRLS
34-
42-
40-
AGE ( mos )
0-5 6-11 12-17 18-23 24-29 30-35 36-41 42-47 48-53 54-59 60+
547
38-
GIRLS
(2)
52-
MALNOURISHED
36 -
50-
TOTAL GROUP
34 -
( Marasmus,
48-
Prekwashiorkor,
Kwashiorkor)
AGE (mos)
T
0-5 6-11 12-17 18-23 24-2930-35 36-41 42-47 48-53 54-59 60+
cm
46-
FIGURE 28 HEAD CIRCUMFERENCE BY SEX AND AGE, MALNOURISHED
44-
CHILDREN (MARASMUS, PREKWASHIORKOR AND KWASHIORKOR)
42.
COMPARED WITH TOTAL GROUP, JORDAN
40-
38
AGE ( mos )
0-5 6-11 12-17 18-23 24-29 30-35 36-41 42-47 48-53 54-59 60+
FIGURE 27 HEAD CIRCUMFERENCES OF MALNOURISHED CHILDREN
COMPARED TO TOTAL GROUP, JORDAN PEDIATRIC STUDY
99
98
ydocum
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
0-50-5 (mos)
Models
4 of 4 columns · click a model to add or remove it

Ours

Exact match

0-5

3 characters4 tokens

Aux 2015

Wrong

We are given a table with data about the head circumference of children in three groups: Marasmus, Prekwashiorkor, and Kwashiorkor. The table shows:

  • Marasmus: 48-
  • Prekwashiorkor: 46-
205 characters64 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

Total Group BOYS Malnourished Boys 48- Prekwashiorkor, Kwashiorkor 46- Marasmus, 48- Prekwashiorkor, Kwashiorkor 42- 48- 52- 48-

128 characters64 tokens