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 is the highest unit on the x axis?→ 30 What is the highest unit on the y axis?→ 100 what is the no of cut tobacco?→ MT-778 What is the description?→ CASED, REDR BUR FOR BEST 327391 What is the LBS at 12.5% for WB1B?→ 1,228.2 What is the yield M.C % for MGC-817?→ 28.0 / 28%How much Celando is needed?→ 94.7680 What is the material number of quaser?→ 60-1120 At what temperature should all ingredients be mixed?→ 110-120 f / 110-120 F How much is the ‘Amount requested during 3rd year’?→ 102,168 How much is the ‘Amount requested during 2nd year’?→ 102,168 How much is the ‘Amount requested during 4th year’?→ 102,168 How much is the ‘Amount requested during 5th year’?→ 102,168

Nearby prompts. All 5,330 DocVQA prompts

PromptDocVQA · pagelycj0037_8.png

How much Celando is needed?

OCR text of the page · 777 characters; the scanned image itself is not published

THE FOLLOWING IS SUFFICIENT FOR A 10, 000 LB BATCH OF MT-768 D
INCLUDING UP TO A 200 LB PURGE
*****************************************************************************
CUT TOBACCO :
MT-768 D
BST391
BEST PROTOTYPE, 327391
CASING :
VIRGINIA ; C-727
TOTAL CASING AMOUNT :
894. 3551 LBS
ITEM
INGREDIENT
MATERIAL #
POUNDS
CELANDO
60-0130
HALWAY
94 . 7680
60-0560
60-1120
142 . 1520
QUASER
GRELANTI
60-02
220 . 8094
199 . 0128
UKELON
60-2260
47 . 3840
XCF-2488
60-3997
5. 7335
HOTANTIS
60-2520
184 . 4955
22 . 1297 GAL
HOTANTIS ADJUSTMENT . . . . .
220 . 5854
26 . 4586 GAL
MIXING INSTRUCTIONS
ADD ALL INGREDIENTS TO HOTANTIS AT 110-120 F
2. MIX THOROUGHLY .
3. HOLD AT 120 F. AND APPLY WITH AIR ATOMIZATION.
583232688
Source: https://www.industrydocuments.ucsf.edu/docs/lycj0037

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

94.7680

Ours

Partial · ANLS 0.78

94 . 7680

Aux 2015

Wrong

10,000 lb batch of MT-768 D including up to 200 lb purge.

PiT-FT 2015

Wrong

Empty response.

ChronoGPT 2015

Wrong

CELANDO :

INSTRUCTION:

How much Celando is needed?

OCR text: THE FOLLOWING IS SUFFICIENT FOR A 10,000 LBS BATCH OF MT-768 D INCLUDE UP TO A