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 · pagehlkb0228_1.png

When was ‘advisory board meeting’ scheduled?

OCR text of the page · 818 characters; the scanned image itself is not published
ADVISORY BOARD MEETING
October 8-10, 1961
ARRIVAL SCHEDULE FOR ADVISORY BOARD MEMBERS AND THEIR WIVES
October 8, Sunday
Dr. and Mrs. Mider
American Airlines Flight 473
Lv Washington
10:15 a. m.
and
Dr. and Mrs. Darby
Lv Nashville
10:55 a. m.
Ar Oklahoma City 1:41 p. m.
Dr. and Mrs. Dunn
American Airlines Flight 22
Lv Los Angeles
8:30 a. m.
Ar Dallas
12:00 noon
DEPARTURE SCHEDULE FOR ADVISORY BOARD MEMBERS AND THEIR WIVES
October 10, Tuesday
Dr. and Mrs. Mider
American Airlines Flight 20
Lv Dallas
4:30 p. m.
Ar Washington
8:55 p. m.
Dr. and Mrs. Darby
American Airlines Flight 274
Lv Dallas
4:35 p. m.
Ar Nashville
7:57 p. m.
October 11, Wednesday
Dr. and Mrs. Dunn
American Airlines Flight 27
Lv Dallas
9:10 a. m.
Ar Los Angeles
10:45 a. m.
(Transportation to and from Ardmore will be arranged by the Foundation)
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
october 8-10, 1961October 8-10, 1961
Models
4 of 4 columns · click a model to add or remove it

Ours

Exact match

October 8-10, 1961

18 characters13 tokens

Aux 2015

Exact match

October 8-10, 1961

18 characters13 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Partial · ANLS 0.89

October 10, 1961

16 characters4 tokens