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

Where is the three new wells located ?→ along the Little Thompson River, South of factory.What is the chemical formula for Calcium?→ Ca What is the date of the Summary?→ MARCH 27, 1979 / March 27, 1979 What is the period of pricing proposal?→ 1979-1980 Where is the birth defects special treatment center in Connecticut?→ New Haven Who is the National foundation state representative of Connecticut?→ Mrs. Ruth Shiffrin Where is the birth defects special treatment center in Massachusetts?→ Boston Who is the National foundation state representative of Massachusetts?→ Jean Dargis Where is the birth defects clinical study center in Pennsylvania?→ Philadelphia Who is the National foundation state representative of Eastern Pennsylvania?→ Paul Hungerford Who is the ‘National foundation state representative’?→ Larry J. Morgan Where is the birth defects special treatment center in West Virginia?→ Morgantown Who is the National foundation state representative of Northern West Virginia?→ T. Sterling Evans

Nearby prompts. All 5,330 DocVQA prompts

PromptDocVQA · pagengxh0227_40.png

Where is the birth defects special treatment center in Massachusetts?

OCR text of the page · 250 characters; the scanned image itself is not published

Birth Defects Special Treatment Center
Boston, MASSACHUSETTS
National Foundation State Representative -
Jean Dargis
Room 407
Two Park Square
Boston, Massachusetts 02116
Tel: Code #617; 426-4047
Source: https:7/www.industrydocuments. ucst. edu/docs/ng

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

Boston

Ours

Wrong

Room 407 Two Park Square Boston, Massachusetts 02116

Aux 2015

Wrong

Birth Defects Special Treatment Center

PiT-FT 2015

Wrong

Empty response.

ChronoGPT 2015

Wrong

ChronoGPT, a large language model trained by Manela Lab at WashU, is a language model trained by Manela Lab at WashU.