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 · pagehqcb0079_29.png

What is the title of the document ?

OCR text of the page · 479 characters; the scanned image itself is not published
SUBSTRATE LOADING
R&D TEST NO.
9-046 J
QUANTITY REQUIRED
150 LBS .
SUBSTRATE TYPE:
98621
INFORMATION :
LOADED FORMULATION:
STEP
1-
Alumina
67,65 8
CAG
10, 00 &
1ST
1ST
3-
FRUCTOSE
2.35 8
1ST
4-
B3
19.94 8
5-
0, 06 &
2ND
FLAV. 98621
2ND
7 -
TOTAL:
100.00 8
8-
9-
FLOW AGENT:
0. 60 8
10-
3RD
11-
12-
SPECIAL INSTRUCTIONS (SAMPLING, TESTING, STORAGE, ETC. ) :
LABEL: ATTN. DAN CHRISTOPHERSON, TRIAL #9-046-J. DATE, RECIPE
SEND SAMPLES FOR ANALYSIS TO QA AND MILHOUS LABS,
51479 8606
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
SUBSTRATE LOADING
Models
4 of 4 columns · click a model to add or remove it

Ours

Exact match

SUBSTRATE LOADING

17 characters6 tokens

Aux 2015

Wrong

SUBSTRATE LOADING

R&D TEST NO.

9-046 J

QUANTITY REQUIRED

150 LBS .

SUBSTRATE TYPE:

98621

INFORMATION :

LOADED FORMULATION:

STEP

1-

Alumina

67,

154 characters64 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

SUBSTRATE LOADING R&D TEST NO. 9-046 J QUANTITY REQUIRED 150 LBS . SUBSTRATE TYPE:

SUBSTRATE TYPE:

Question:

What is the title of the document ?

151 characters64 tokens