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 · pagezlkd0079_3.png

What is the first word on the page?

OCR text of the page · 2,346 characters; the scanned image itself is not published
years experience, it is prudent to cover at least one of these in the Toxicology
Research's 9010 budget.
Neuropharmacology
Dr. J. Gray
99.2
146.0
Research
Univ. of London
A battery of behavioral tests will be used to study the effects of nicotine on
reference and working memory in normal, anosmic, and cholinergically-deficient rats.
Program Consulting
To Be Determined
346.1
415.0
105
Pesticides (Stemmery)
Hazleton
315.0
300.0
Conduct pesticide residue analyses on 1988 domestic tobacco (burley & flue cured)
purchased by RJRT. 1987analyses revealed the use of endosulfan and maleic hydrazide.
Analyses needed for 3-4 years to build database on pesticide use and levels by U.S.
farmer.
Pesticides (Brands)
Hazleton
300.0
250.0
Evaluate pesticide levels on tobaccos used for new brands, e.g. VRP, and on NPT
tobacco, e.g. CAG per tobacco plug.
SARA Documentation
250.0
and MSDS Database
Ruger isIt Corporation
0.0
Prepare a-computer database which allows preparation of Environmental Protection
Agency and Occupational Safety and Health Administration reporting forms, acess by
employees to material safety data sheets and company-wide hazardous chemical storage
locations. Current manual proceedures cost $150,000 per year (estimated). Annual
maintenance of program to cost $20,000 serviced by RJRT Iformation Resources.
Evaluation of Chemicals
SRI
125.0
150.0
Contract required to provide documentation of chemical reviews as needed by R&D.
Examples include solvents used by PT&D, paper components for cigarette papers, and
materials for NPT.
Freon Replacement
To Be Determined
1500.0
Generate sufficient chemical and biological database to support the use of Freon-11
replacements. Currently PT&D expects to evaluate two Freons and propane. Costs
include $100,00 for genetic toxicology work; $500,00 for 90-day inhalation study;
$300,000 for chemical analyses, and $600,000 to begin chronic skin painting work in
1989. Partial costs may be recouped if industry plan is developed to share costs.
Lorillard has agreed to this concept.
Animal Research
To Be Determined
500.0
(Nutrition)
Allocated by Nabisco Brands
Nabisco Brands has requested us to supervise the toxicological research of new food
additives at a contract laboratory. Costs will be refunded to RJRT by NBI.
Consultants for RJRTI-
To Be Determined
100.0
NPT Project
2
50823 1373
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
years
Models
4 of 4 columns · click a model to add or remove it

Ours

Wrong

Research's

10 characters3 tokens

Aux 2015

Wrong

The first word on the page is "years".

38 characters11 tokens

PiT-FT 2015

Wrong

<|<assistant|> <|<assistant|> <|<assistant|> <|<assistant|> <|<assistant|> <|<assistant|> <|<assistant|> <|<assistant|>

120 characters64 tokens

ChronoGPT 2015

Wrong

Nabisco Brands Nabisco Brands has agreed to this concept. Costs will be refunded to RJRT Iformation Resources.

Question:

What is the first word on the page?

Nabisco Brands Nabisco Brands has agreed to this concept. Costs

227 characters64 tokens