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 · pageqxpn0081_9.png

What were the two Menthol Load types used for test?

OCR text of the page · 1,019 characters; the scanned image itself is not published
KAYMICH III - PRINT/INK TYPE, COLOR/PAPER TYPE/
MENTHOL LOAD TEST
VARIABLES:
PRINT
DIAMOND OR SALEM GOLD LOGOS
INK
GREEN OR GOLD
PAPERS
REF. 856, REF. 456, REF. 856 HV, 856 AQUAPEL, 45 G
CACO, FILLER PAPER, ZZ PAPER 50 G MG(OH), WITH KCL
MENTHOL -
.12% OR .32%
LOADS
TEST ACCOMPLISHED WITH SIX 2 X 2 DESIGNS - 24 PARTS
ALL PARTS RAN IN SPC
ALL PARTS PACKAGED IMMEDIATELY OFF MAKER/TIPPER (WITHIN 15
MINUTES)
ALL TEST LEGS WITHIN EACH 2 X 2 WERE NOT SIGNIFICANTLY
DIFFERENT* IN TOBACCO AND SMOKE ANALYSES
OBVIOUS ANALYTICAL SMOKE DIFFERENCES WERE PRESENT
BETWEEN DESIGNS OF DIFFERING PAPER TYPE
TWO TRENDS EXISTED
1. MENTHOL LOAD FROM KAYMICH TRENDED LOW
2. FILTER MENTHOL FROM KAYMICH TRENDED HIGH
SED DISCRIMINATION OF INTERNATIONAL PRODUCTS AND SALEM
GOLD PROTOTYPES WERE NOT SIGNIFICANTLY DIFFERENT. SALEM
BOX PROTOTYPES WERE DIFFERENT. THESE WERE SUBMITTED FOR
QDA ANALYSIS.
SED DIAGNOSTIC ANALYSES OF SALEM BOX PRODUCTS INDICATED
NO PERCEPTUAL TASTE DIFFERENCES. VISUAL DIFFERENCES IN
PRINT COLOR, ONLY.
51335 8998
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
.12% OR .32%.12% or .32%
Models
4 of 4 columns · click a model to add or remove it

Ours

Exact match

.12% OR .32%

12 characters10 tokens

Aux 2015

Wrong

KAYMICH III - PRINT/INK TYPE, COLOR/PAPER TYPE/ MENTHOL LOAD TEST VARIABLES: PRINT DIAMOND OR SALEM GOLD LOGOS INK GREEN OR GOLD PAPERS REF. 856, REF. 456,

155 characters64 tokens

PiT-FT 2015

Wrong

Question: What were the two Menthol Load types used for test?

OCR text: KAYMICH III - PRINT/INK TYPE, COLOR/PAPER TYPE/ MENTHOL LOAD TEST VARIABLES: PRINT DIAMOND

164 characters64 tokens

ChronoGPT 2015

Wrong

KAYMICH III - PRINT/INK TYPE, COLOR/PAPER TYPE/ INK

Question:

What were the two Menthol Load types used for test?

OCR text: KAYMICH III - PRINT/INK TYPE, COLOR/PAP

169 characters64 tokens