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 · pagerlvd0227_2.png

Which is the manuscript 'in possession'?

OCR text of the page · 886 characters; the scanned image itself is not published
May 3, 1968
MANUSCRIPT COMMITTMENTS
Author
Subject
Date
Manuscript Due
Thomas Brewer, M.D.
Nutritional Aspects in Eclampsia
4/68
5/68
Frances Berkeley
Little Land Snail
3/68
In possession
Floore
William Darby, M.D.
Nutritional Status of Man
1/68
5/ 1
Dr. F. W. Hill
Programming Nutrition -- Animals
3/68
7 / 1
Sir Raymond Priestly
An Interview with Sir Raymond Priestly
5/68
5/ 68
Dr. William Siri
Nutrition and Mountain Climbing
3/68
7/1/68
Capt. Villiers:
An Afternoon with Alan Villiers
4/68
The explorer and his food)
5 / 68
Samuel Lepkovsky, Ph. D.
Taste
1967
6/15/68
Franz J. Ingelfinger, M.D.
The Large Bowel
12/67
7/68
Lowell Thomas
Willard Krehl, M.D.
Nutrition and the Explorer
11/67
5/14
SUBJECTS UNDER DISCUSSION
C.B. Brown, M.D.
Hospital "Eat-In"
4/68
No reply as yet.
Mrs. John P. Conrad
Food Allergies
2/68
Would like to
do article.
.industrydocuments.ucst.edu/docs/rival
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
Little Land Snail
Models
4 of 4 columns · click a model to add or remove it

Ours

Exact match

Little Land Snail

17 characters5 tokens

Aux 2015

Wrong

The manuscript 'in possession' is the manuscript 'in possession'

64 characters14 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

Manuscripts due to be published in the next few months.

Question: Which is the manuscript 'in possession'?

OCR text: Manuscripts due to be published in the next few months.

Manuscripts due to be published in the next few months

231 characters64 tokens