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 · pageysmc0228_6.png

What is the percentage of men in 40-49 age group exceeding best weight by 10% or more?

OCR text of the page · 410 characters; the scanned image itself is not published
TOO MANY PEOPLE ARE OVERWEIGHT
Proportion Exceeding "Best" Weight
By 10% or more
By 20% or more
Age
When
Insured
MEN
31 %
20-29
12%
53%
30-39
25%
60%
40-49
32%
63%
50-59
34%
WOMEN
23%
20-29
12%
41%
30-39
25%
59%
40-49
40%
67%
50-59
46%
"Best" weight is slight or moderate underweight, ranging
from about 5-10% below average weight at age 30 to
about 10-15% below at ages 45 and over.
Source:
.edu/does/ysmc0228
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
60%
Models
4 of 4 columns · click a model to add or remove it

Ours

Wrong

32%

3 characters4 tokens

Aux 2015

Wrong

To find the percentage of men in the 40-49 age group exceeding "best" weight by 10% or more, we first need to determine the "best" weight range. The OCR text states that "best" weight is "slight or moderate underweight, ranging from about 5

240 characters64 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

The percentage of men in 40-49 age group exceeding best weight by 10% or more.

78 characters19 tokens