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 · pageknmk0226_2.png

What is the telephone number given at the bottom of the ‘letter’?

OCR text of the page · 1,357 characters; the scanned image itself is not published
File
Carl Byoir & Associates, Inc.
MAR 1 3 1978
Carolyn Walden, Director
TV. Radio, Film Department
at omen Bill . brindaud yo tod
boy base of sat anow bluFebruary 9, 1978
aid bandand quoy
Ms. Julia Child
"The French Chef"
clo WGBH-TV
125 Western Avenue
Boston, Massachusetts 02134
Dear Ms. Child:
Boy. ...do you have a lot of fans! Unfortunately for me, a lot of
them are members of the Sugar Association which is a client of mine.
I say unfortunately because in the last month or so it seems I've
heard from each and every one of the Association's members.
The reason? It seems in a recent broadcast you mentioned that you
were about to use cane sugar as an ingredient. . . instead of beet
sugar. ..because of its superior qualities. Well, as you'll soon
see from the enclosed materials, the authorities say there is no
difference between the two. That, indeed, sugar -- whether refined
from sugar beets or from sugar cane -- is the same sugar. I guess
(to paraphrase Gertrude Stein) you could say that "sugar. ..is sugar
. ..is sugar."
This should serve as an explanation for all that I've sent. I hope
you find the material both interesting and informative. And if I
can be of any additional help, please don't hesitate to call me
collect in New York.
Cordially,
Carolyn Wilder
(more)
380 Madison Avenue, New York, N. Y. 10017 . Telephone (212) 986-6100
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
(212)986-6100
Models
4 of 4 columns · click a model to add or remove it

Ours

Partial · ANLS 0.93

(212) 986-6100

14 characters15 tokens

Aux 2015

Wrong

380 Madison Avenue, New York, N. Y. 10017. Telephone (212) 986-6100

67 characters36 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

file chrno chrno file file file

31 characters15 tokens