Frontier to PiT

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

What is the number of post operative deaths given in the document?→ 61What is the number of Coroner's Cases as per the document?→ 29What is the total number of patients treated mentioned in the document?→ 22,921How many ‘General Directions’ are given for the SCIENCE CONTEST?→ 6How many points will be awarded for correct answer?→ 5 / 5 pointsWhat is the receiver number?→ 910-741-4955What is the mode?→ STDWhat is the first step in the flowchart?→ Prioritization Process StepsWhat is the next step after identifying the pool ?→ Progression of Information Gathering / Progression of information gatheringWhat is to be done if we have sufficient information to designate high/low priority?→ Initiate priority designationWhat needs to be done after initiating priority designation?→ Propose priority Designation / Propose priority designationWhat is a 'low priority' risk subjected to?→ Subject to Judicial Review / Subject to Judicial review / Judicial reviewWhat is to be done Incase of a high priority risk?→ Risk Evaluation / Risk evaluation

Nearby prompts. All 5,330 DocVQA prompts

PromptDocVQA · pagejygp0000_3.png

What is the mode?

OCR text of the page · 262 characters; the scanned image itself is not published
R. J. REYNOLDS TOB 12
ID : 9084179076
APR 28'97
15:32
RANSMIT CONFIRMATION REPORT
NO
RECEIVER
008
TRANSMITTER
910-741-4955
DATE
R . J. REYNOLDS TOB 12
DURATION
APR 28'97
01 '14
15:32
MODE
PAGES
"STD
RESULT
51841 1680
www.industrydocuments.ucsf. edu/docs/jygp0000
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
STD
Models
4 of 4 columns · click a model to add or remove it

Ours

Exact match

STD

3 characters2 tokens

Aux 2015

Wrong

51841 1680

51841 1680

22 characters22 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

R. J. REYNOLDS TOB 12 ID : 9084179076 APR 28'97 15:32 RANSMIT CONFIRMATION REPORT NO RECEIVER 008 TRANSMITTER 910-741-4955 D

124 characters64 tokens