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 · pagemfnf0004_2.png

What is the common agenda for both the days?

OCR text of the page · 349 characters; the scanned image itself is not published
Agenda
Wednesday, July 17
10:00
Welcome Speech (Mr. TADA)
10:05
Report on Europe and US launch (Mr. Bass)
11:00
Launch Plan for NSP Product (Mr. Fujimoto)
12:00
Launch
13:00
License Negotiation
16:30
Adjourn
18:30
Dinner
Tuesday, July 18
10:00
License Negotiation
12:00
Adjourn
51722 0577
Source: https://www.industrydocuments.ucsf.edu/docs/minf0004
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
License Negotiation
Models
4 of 4 columns · click a model to add or remove it

Ours

Exact match

License Negotiation

19 characters4 tokens

Aux 2015

Wrong

Agenda Wednesday, July 17 10:00 Welcome Speech (Mr. TADA) 10:05 Report on Europe and US launch (Mr. Bass) 11:00 Launch Plan for NSP Product (Mr. Fujimoto) 12

157 characters64 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

Tuesday, July 18 10:00 Report on Europe and US launch (Mr. TADA) 11:00 Launch Plan for NSP Product (Mr. Bass) 12:00 Launch 16:30 Dinner Tuesday, July 18 10:00 Report on Europe and

179 characters64 tokens