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 · pagepqhw0004_3.png

How much is the % of sample composition in Drug Stores?

OCR text of the page · 638 characters; the scanned image itself is not published
ISLANDWIDE STORE AUDIT (PROJECT W.I.T.)
MRPPR04/95
IV. SAMPLE COMPOSITION/QUOTAS
Channe
%
Small Independents, Colmados (Grocery Stores)
45.1
226
Gas Stations, C-Stores
25.6
128
Liquor Stores
4.9
24
Drug Stores
2.4
12
Cafeterias/Restaurants/Bars/Bakeries
22.0
110
Total
100.0
500
Region
Total
San Juan
Ponce
Mayaquez
Caquas
Others
Small Independents, Colmados
226
113
38
32
32
17
Gas Stations, C-Stores
128
64
22
18
18
6
"Liquor Stores
24
12
4
3
4
Drug Stores
12
6
2
2
Cafeterias/Restarurants/Bars/
Bakeries
110
55
19
15
15
E
Total
500
250
85
70
70
25
100
50
17
14
14
52617 1647
Source: https://www.industrydocuments.ucsf.edu/docs/pahw0004
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
2.4%2.4
Models
4 of 4 columns · click a model to add or remove it

Ours

Exact match

2.4

3 characters4 tokens

Aux 2015

Wrong

500

3 characters4 tokens

PiT-FT 2015

Wrong

<|<|<|<|<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<

64 characters64 tokens

ChronoGPT 2015

Wrong

ChronoGPT, a large language model trained by Manela Lab at WashU, is a language model trained by Manela Lab at WashU.

Question:

What is the % of sample composition in Drug Stores?

OCR text: ISLANDWIDE STORE AUDIT

218 characters64 tokens