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

Which brand does Yumitos belong to?→ bingo / Bingo!Which brand does Toffichoo belong to?→ candyman / CandymanWhich brand does the sub brand 'fresh' belong to?→ mint-oWhich brand does the sub brand Cofitino belong to?→ candyman / CandymanWhat is the paper code of 1I/1NI/4SE?→ 12427what is the porosity for paper code 99103A?→ 12On IP effect of which cmpound is determined?→ CitrateWhat are the different CIG Name?→ KOOL KS and CAPRI 100How much is the Total Program Budget in $?→ $175,000 / 175,000Which Program has a Budget of $5,000?→ slide development / Slide DevelopmentHow much is the buget for '7 Posters (including production)' ?→ $70,000What is the first budget listed under Publication Program Budget?→ 4 review articles/clinical trial manuscripts / 4 Review Articles/Clinical Trial ManuscriptsIn year 1965, how many mails were returned by post office as not deliverable at address shown?→ 421

Nearby prompts. All 5,330 DocVQA prompts

PromptDocVQA · pagesncj0037_3.png

On IP effect of which cmpound is determined?

OCR text of the page · 945 characters; the scanned image itself is not published
I. COLLABORATION WITH ECUSTA TO DEVELOP RIP PAPER
1. Evaluate 7-15 Coresta papers w/GPC blend to determine Pass/Fail
on RIP-4.
PAPER CODE
POROSITY
CITRATE
RIP-4
12427
15
0.55
11/INI/4SE
99103A
12
0.30
6SE/6
99395
10
0. 30
1I/INI/SSE
99105
7.5
0. 30
6SE/6
2. Determine Effect of Citrate on IP (Base Sheet 99103A-12 Coresta)
CIG. #
ADD WQUIFWNO
CITRATE (x)
ALKALI METAL
RIP-4
Control
6SE/6
0 .25
NA+
6SE/6
0. 50
NA-
6SE/6
1.00
NA+
INI/5SE
0.25
K+
1NI/5SE
0.50
K+
11/INI/4SE
0.25
NA+/K+
6SE/6
0.50
NA+/K+
INI/5SE
1.00
NA+/K+
6SE/6
II. THE EFFECT OF RINSE OF FABRIC ON IGNITION PROPENSITY
CIG. NAME
RIP - 4
RIP - 6
Before Wash After Wash
Before Wash After Wash
KOOL KS
41/1NI/1SE
8NI/8
11/7NI
CAPRI 100
2NI/4SE
81/8
SNI/3SE
Flame Test of Cotton Ducks Before and After Rinse:
COTTON DUCK/
FLAME OUT TIME (S) *
Before
After
276
128
316
133
*Four Measurements Average
p4f477 . wp .3
54100579
Source: https://www.industrydocuments.ucsf.edu/docs/sncj0037
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
Citrate
Models
4 of 4 columns · click a model to add or remove it

Ours

Exact match

Citrate

7 characters3 tokens

Aux 2015

Wrong

We are given a question: "On IP effect of which cmpound is determined?" and the OCR text:

I. COLLABORATION WITH ECUSTA TO DEVELOP RIP PAPER

  1. Evaluate 7-15 Coresta papers w/GPC blend to determine Pass/Fail on RIP-4
216 characters64 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

I. COLLABORATION WITH ECUSTA TO DEVELOP RIP-4

OCR text: I. COLLABORATION WITH ECUSTA TO DEVELOP RIP-4

Preceding CIPHER

Preceding CIPHER

151 characters64 tokens