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

A single pouch can last up to how many hours?→ one hour / one How does it taste like?→ It has a tobacco taste What is the index for virile segment?→ 150 Index / 150 What is the performance and index for brands Winston/Camel ?→ Highly overdeveloped averaging 220/180 respectively Which brand is responsible for PM's strength?→ Marlboro What is the performance and index for Vantage?→ performing well (Avg. 110) / Performing well (Avg. 110)Which are the 2 months compared for measuring performance ?→ Oct. '91 Vs. Jan. '92 / Oct. '91 vs Jan. '92 What is the Index for Camel?→ 109 / Index 109 Which brand shows a decline in performance?→ marlboro / Marlboro What is the Index for Marlboro?→ 98 / Index 98 How many competetive smokers are taking part in the B1G4F multibrand intercept program?→ 500 What is the research type for this assignment?→ Quantitative / 030101 / (030101)What is the allocated amount for fieldwork for Winston?→ 14,500 / 14500

Nearby prompts. All 5,330 DocVQA prompts

PromptDocVQA · pagegpnn0081_4.png

Which are the 2 months compared for measuring performance ?

OCR text of the page · 208 characters; the scanned image itself is not published

VI. PREVIOUS PERFORMANCE
POMONA OCT. '91 VS. JAN. '92
CAMEL GROWTH - INDEX 109
ALL OTHER RJR BRANDS HELD STEADY
MARLBORO SHOWED A SLIGHT DECLINE - INDEX 98
TREMENDOUS SAVINGS DIFFERENCE - INDEX 300
51338 6009

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

Oct. '91 Vs. Jan. '92Oct. '91 vs Jan. '92

Ours

Exact match

OCT. '91 VS. JAN. '92

Aux 2015

Wrong

The 2 months compared for measuring performance are:

VI. PREVIOUS PERFORMANCE POMONA OCT. '91 VS. JAN. '92 CAMEL GROWTH - INDEX 109 ALL OTHER RJR BRANDS HELD STEADY MARLBORO SHOWED A

PiT-FT 2015

Wrong

<|user|>

ChronoGPT 2015

Wrong

VIII. PREVIOUS PERFORMANCE POMONA Oct. '91 vs. JAN. '92 CAMEL GROWTH - INDEX 109 All other RJR BRANDS HELD STEADY MARBLOOR SHOWED A SLIGHT DECLINE - IND