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 · pagefgbd0079_4.png

What is the PD ?

OCR text of the page · 210 characters; the scanned image itself is not published
SENT BY:RJ REYNOLDS TOB. CO. ; 8- 9-94 ; 15:25 ;PT&D MECH & INST DEV-
9197410918;# 4
. .: .
CSF Run Sheet
Date:
PD- 5254 8
CSF- 2-2Z
Run Nos.
1883 - 1897
Net Pounds Infeed:
893
Net Pounds Out:
770
86
51441 8651
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
5954 B
Models
4 of 4 columns · click a model to add or remove it

Ours

Partial · ANLS 0.67

5254 8

6 characters7 tokens

Aux 2015

Wrong

PD- 5254 8

CSF- 2-2Z

Run Nos.

1883 - 1897

Net Pounds Infeed:

893

Net Pounds Out:

770

86

51441

103 characters64 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters1 tokens

ChronoGPT 2015

Wrong

PD- 5254 8 PD- 2-2Z Run Nos. 1883 - 1897 Net Pounds Infeed: 893 Net Pounds Out: 740 86 51441 8651

97 characters45 tokens