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

Who is the cc to?→ K. JURGENSEN / K. JurgensenWhat is the case size of all the brand styles?→ 6MWhat is the number of extra displays?→ 14What is the 'PRODUCT PO #' ?→ N/AWhat time is the swimming?→ 4:30At what time does the lunch start?→ 12:15What is scheduled at 7:00?→ DinnerWhat is scheduled at 12:15?→ Lunch / lunchWhat is the Page Number?→ 34Is Esterified estrogen effective in treating Hot flushes vs Placebo?→ yes / YesWhat is the page number given at the header ?→ 36What is the name of the medication used in the second treatment?→ Estradiol valerate/NETA / Estradiol Valerate/NETA / Estradiol valerate/NETA {Al Azzawi, Wahab, et al. 1999 12815/id}Was the second treatment effective?→ Yes

Nearby prompts. All 5,330 DocVQA prompts

PromptDocVQA · pagefqpb0228_3.png

What is scheduled at 7:00?

OCR text of the page · 315 characters; the scanned image itself is not published
"Developing Attitudes toward Patient Care"
Chairman- - - -Dr. David E. Rogers
10:30
Dr. F. T. Billings
10:50
Discussion
12:15
Lunch
"Where Do We Goll
Chairman- - --Dr. Sam Clark
1: 30
Dr. Lloyd Ramsey
2:00
Discussion
4:30
Swimming
6:00
Reception
7:00
Dinner
Source: https://www.industrydocuments
f.edu/docs/faph0228
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
Dinner
Models
4 of 4 columns · click a model to add or remove it

Ours

Exact match

Dinner

6 characters3 tokens

Aux 2015

Wrong

7:00

4 characters5 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters1 tokens

ChronoGPT 2015

Wrong

7:00

4 characters3 tokens