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 · pagejpgn0226_221.png

What is the first point under 'discussion points'?

OCR text of the page · 1,171 characters; the scanned image itself is not published
Attendees:
Investigators
DePuy
John Callaghan, M.D.
John Blachere
Brian Haas, M.D.
Nancy Friddle
Wayne Paprosky, M.D.
Steve Haas
Tom Schmalzried, M.D.
Jim Lancaster
Steve Wentworth
Meeting Topics List
Item
Section
Discussion Points
. . .
Purpose/Objectives
Protocol;p. 2
Can we prove them?
Inclusion/Exclusion Criteria
Protocol; pp. 5-7
Review criteria
Success/Failure Criteria
Protocol; pp. 11-13
Review criteria
Patient Follow-up
Protocol; pp. 9-10
Clinical/X-ray intervals appropriate?
Radiographic Technique
Radiographic Protocol
Suitable technique?
Hip Rating System
Protocol; p. 10
Harris Hip, SICOT, other?
Case Report Forms
Case Report Forms
Format and content (Harris, SICOT,
and 2 different x-ray forms included)
Risk Analysis
Risk Analysis
All reasonable risks associated with
device.
Other Topics
Independent Radiologist- Do We Need One?
Serum and Urine Lab Tests- What Kinds? How Often?
Poly Ware Software- J. Blachere to discuss
IDE Investigator Obligations- N. Friddle to discuss
R.USERS\SJW\1997\MOMDR.LET
PROTECTED DOCUMENT. DOCUMENT SUBJECT TO PROTECTIVE ORDER.
DEPUY086231695
Source: https://www.industrydocuments.ucsf.edu/docs/jpgn0226
PLT-03045-00613
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
Can we prove them?
Models
4 of 4 columns · click a model to add or remove it

Ours

Partial · ANLS 0.56

Protocol;p. 2 Can we prove them?

32 characters11 tokens

Aux 2015

Wrong

The first point under "Discussion Points" is the first point under "Discussion Points" in the document.

103 characters22 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

OCTOBER, 1998 OCTOBER, 1998 OCTOBER, 1998 OCTOBER, 1998 OCTOBER, 1998 OCTOBER, 1998 OCTOBER, 1998 OCTOBER, 1998 OCTOBER, 1998 O

127 characters64 tokens