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 · pagehjbl0037_5.png

What is the first point under 'developing'?

OCR text of the page · 1,146 characters; the scanned image itself is not published
NICOTINE RSM
1921 COMPLETION MILESTONES
1991 RESEARCH PROJECTS
1991-1995 CONTINUED RESEARCH
Completion of all test methods for
.
Begin testing/implementation-
Determine optimal nicotine, 'tar',
test-methods for-diagnostic learning
of 'biggest ticket' product/
draft among homogeneous smoker
of smoker satisfaction fler ( 2 Q)
concept
groups (product and/or mindset).
Complete Wave II HSB
Begin ULT Nicotine Optimization
. Implement improved products to
Axial points
Program
the marketplace with the first
-
Optimization of best nicotine, 'tar', draft
implementation in 1991. /992-
level among specified mindsets of FFLT/FF
.
Determine need for REST tobaccos for
smokers (Q) (4Q)
nicotine modifications
Design and conduct similar study
Final results/recommendations (2090) (/69/)
with focus on ULT improvement.
DRIVING SCIENCE/TECHNOLOGIES
EXISTING
DEVELOPING
NEEDED
HMSM
SULT products
Extraction processes
DFC
.
REST tobaccos/nicotine
GLA tobaccos
.
Plasma nicotine estimates
extraction processes
HSB
New consumer groups defined by
. HMSM
psychographic/lifestyle information
51092 5215
Source: https://www.industrydocuments.ucsf.edu/docs/hib10037
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
sult productsSULT Products
Models
4 of 4 columns · click a model to add or remove it

Ours

Wrong

REST tobaccos/nicotine

22 characters8 tokens

Aux 2015

Wrong

The first point under 'developing' is the first point under 'developing' in the text.

85 characters22 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

NICOTINE RSM 1921 COMPLETION MILESTONES 1921-1922 CONTINUED RESEARCH 1921-1922 CONTINUOUS RESEARCH 1921-1922 CONTINUOUS RESEARCH 1921-1922 CONTINUOUS RESEARCH 19

161 characters64 tokens