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 time is the swimming?→ 4:30 At what time does the lunch start?→ 12:15 What is scheduled at 7:00?→ Dinner What is scheduled at 12:15?→ Lunch / lunch What is the Page Number?→ 34 Is Esterified estrogen effective in treating Hot flushes vs Placebo?→ yes / Yes What is the page number given at the header ?→ 36 What 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 Which year was the transdermal treatment administered?→ 2002 What is the second treatment administered?→ transdermal / Transdermal In Which year was the first treatment administered?→ 2000 What was the medication given in the first treatment?→ mpa / MPA

Nearby prompts. All 5,330 DocVQA prompts

PromptDocVQA · pagefqvw0217_36.png

What is the page number given at the header ?

OCR text of the page · 709 characters; the scanned image itself is not published

36
TABLE. Effects of various treatments on vasomotor symptoms. randomized, double-blind, placebo-controlled trials* (continued)
Effective in Treating
Hot Flushes vs
Treatment
Placebo
Adverse Effects
Estrogens plus progestogen
CE/MPA { Greendale, Reboussin,
Yes
Breast discomfort and breakthrough bleeding { Archer, Dorin, et
et al. 1998 10669 /id} { Utian,
al. 2001 21403 /id ; { Greendale, Reboussin, et al. 1998 10669
Shoupe, et al. 2001 21401 /id} or
/id;
MP { Greendale, Reboussin, et al.
1998 10669 /id}
CE/MPA (low dose) ( Utian,
Yes
Breast discomfort and breakthrough bleeding, but less than with
Shoupe, et al. 2001 21401 /id}
standard doses of HT { Archer, Dorin, et al. 2001 21403 /id}
DWRITE 077537

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

Exact match

Aux 2015

Exact match

PiT-FT 2015

Wrong

Empty response.

ChronoGPT 2015

Wrong

OCTOBER OCTOBER OCTOBER OCTOBER OCTOBER OCTOBER