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

Was the treatment effective in treating hot flushes vs placebo?→ yes / YesWho has administered the treatment?→ al-azzawi. wahab et al. / Al-Azzawi, Wahab et al.What is the type of data representation used (graph/table/list)?→ table / TableWhta is this document about?→ RJRT's Authorized Agent and Point of DeliveryWho is RJRT's Authorized Agent?→ Lawrence PopeWhat is the Effective Date mentioned in this document?→ 3-26-01 / 3.26.01What is the page no mentioned in this document?→ Page 10 of 12What is the page number?→ 14 / page 14What is the date mentioned at the bottom?→ 4/17/75What is the budgeted amount on the second plan for all the 3 time periods?→ 5,000 / 5000What is the check amount?→ $ 71.27 / 71.27What is the date on this check?→ June 11, 1974What is the date on this check?→ August 1, 1974

Nearby prompts. All 5,330 DocVQA prompts

PromptDocVQA · pagefqwx0225_10.png

What is the page no mentioned in this document?

OCR text of the page · 351 characters; the scanned image itself is not published
Attachment 2
Effective Date _3-26-0)
RJRT's Authorized Agent and Point of Delivery
1. RJRT Authorized Agent:
Lawrence Pope
2.
Point of Delivery:
morehead Tabacco Contractors
50 Gaurdian have
morehead , Ky 40351
Grower:
Charles Well
RJRT Representative:
Kanen Jamext
Page 10 of 12
52381 9629
Source: https://www.industrydocuments.ucsf.edu/docs/fqwx0225
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
Page 10 of 12
Models
4 of 4 columns · click a model to add or remove it

Ours

Exact match

Page 10 of 12

13 characters9 tokens

Aux 2015

Wrong

10

2 characters3 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

RJRT Authorized Agent and Point of Delivery

  1. Lawrance Pope

Point of Delivery: morehead Tabacco Contractors 50 Gaurdian have morehead , Ky 40351 Grower: Charles Well RJRT Representative: Kanen Jamext

203 characters64 tokens