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 · pagezzcw0217_10.png

Who reviewed Bone paper?

OCR text of the page · 938 characters; the scanned image itself is not published
Minutes From the Menopausal Health Publication Management Committee Meeting
10
January 16, 2002
VI. Women's HOPE Publications
Bone paper is being reviewed by JAMA
Analysis of bleeding review article will be submitted to Climacteric
VII. Continuing Medical Education
American Medical Women's Association (January 31-February 3; San Antonio, TX):
HRT breakfast symposium presented by Drs. Richard-Davis and Wallace
Two symposia will be held at IMS, one Premarin-sponsored (advocating low-dose-
Lindsay, Fiorica, Rosano, Flaws to present), one Totelle-sponsored
VIII. PR Opportunities
. Ketchum pursuing proclamation re: Trudy Bush's work with Maryland's state
legislature
Potential press for Civitelli's article on alveolar bone loss
Potential interview with Clarkson on his research (to spread positive MPA message)
Confidential Pursuant to Confidentiality
CONTA025-023691
Order
Source: https://www.industrydocuments.ucsf.edu/docs/zzcw0217
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
JAMAJama
Models
4 of 4 columns · click a model to add or remove it

Ours

Exact match

JAMA

4 characters3 tokens

Aux 2015

Wrong

Minutes From the Menopausal Health Publication Management Committee Meeting

10 January 16, 2002 VI. Women's HOPE Publications Bone paper is being reviewed by JAMA

Analysis of bleeding review article will be submitted to Climacteric

VII. Continuing Medical Education American Medical

285 characters64 tokens

PiT-FT 2015

Wrong

Question: Who reviewed Bone paper?

OCR text: Minutes From the Menopausal Health Publication Management Committee Meeting 10 January 16, 2002 VI. Women's HOPE Publications Bone paper is being reviewed by JAMA Analysis of bleeding review article will be submitted to Climacteric

280 characters64 tokens

ChronoGPT 2015

Wrong

ChronoGPT, a large language model trained by Manela Lab at WashU, is reviewing Bone paper.

Question:

Who reviewed Bone paper?

OCR text: Minutes From the Menopausal Health Publication Management Committee Meeting 10 January 16, 2002 VI. Women

247 characters64 tokens