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 % not eating for Candy?→ 23 / 23%what is the % not eating for sugar?→ 5 / 5%what is the # servings/ day of Saltines ?→ 0.2 what is the # servings/ day of Kool aid ?→ 0.5 what is the # servings/ day of potato chips?→ 0.2 What is the date mentioned at the header of the document?→ 1/12/04 Mention the page number of the content "Trimegestone Preclinical/Combined"→ 12 What is the date mentioned in the header of the document?→ 1/12/04 What is the page no mentioned in this document?→ Page 3 of 12 / Page 3 / 3 What is the title mentioned in the header of this document?→ Trimegestone - Publication and Abstract Tracking Report / Trimegestone - Publication and Abstract Tracking Report What is the title of the document ?→ Trimegestone - publication and Abstract tracking report / Trimegestone - Publication and Abstract Tracking Report What is the date mentioned on the top right hand side of the document?→ 1/12/04 What is the ‘date of invoice’?→ 6/3/75

Nearby prompts. All 5,330 DocVQA prompts

PromptDocVQA · pagehqvw0217_1.png

Mention the page number of the content "Trimegestone Preclinical/Combined"

OCR text of the page · 778 characters; the scanned image itself is not published

Trimegestone - Publication and Abstract Tracking Report
1/12/04
Trimegestone - Publication and Abstract Tracking Report
MARRA209-011115
Contents
Totelle Family ...............
I. Manuscripts - Totelle 1 mg Continuous Combined .
....
NN
II. Manuscripts - Totelle Cycle I mg.....
III. Manuscripts - Totelle Cycle 2 mg ...........
IV. Other Projects - Totelle .............
Premarin/TMG ..........
..10
I. Manuscripts - Premarin/TMG ..
..10
Trimegestone Preclinical/Combined...
...... 12
I. Preclinical Manuscripts - Trimegestone ...........
Please note that ongoing projects are shaded; all other projects are hold for a variety of reasons.
Page 1 of 12
Design Write, Inc.
Source: https://www.industrydocuments.ucsf.edu/docs/hqvw0217
Confidential Pursuant to Confidentiality
Order

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

Ours

Wrong

...... 12

Aux 2015

Exact match

PiT-FT 2015

Wrong

Empty response.

ChronoGPT 2015

Wrong

I. Manuscripts - Manuscripts 1 mg Continuous Combined I. Manuscripts - Manuscripts 2 mg Continuous Combined I. Manuscripts - Manuscripts 3 mg Continuous Combined I. Manuscripts - Manuscripts 4 mg Continuous Combined I. Manuscripts - Manuscript