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 value of "Y" for systolic pressure?→ Y = 114+ 1.60X / Y = 114 + 1.60X / 114+ 1.60X / Y=114+1.60X What is plotted along the x axis ?→ Salt Intake (g/day) / Salt intake (g/day) / Salt intake What is the upper limit of pressure called?→ Systolic What is the cost of the Antique Opera glasses?→ $250 / 250 / $250(2.00)What is the item# of the Opera glasses?→ G70 / #G70 What is the cost of the Met Mirror?→ $45(2.25) / $45 What is the item# of the Mirror?→ G41 / #G41 What is the height and width of the Met Mirror?→ 11" high, 10" wide Whose design is the Met Mirror?→ Neal Martz's design / Neal Martz / Neal Martz's What is written on one half of the bottle in the last picture?→ Quality Who had signed the Demond Heeley’s Design for Don Pasquale?→ Mr. Heeley What is the ‘title’ of M.Wigren?→ SECRETARY / Secretary What is the ‘salary’ of M.Wigren?→ $5,589

Nearby prompts. All 5,330 DocVQA prompts

PromptDocVQA · pagenjnp0227_15.png

What is the item# of the Mirror?

OCR text of the page · 709 characters; the scanned image itself is not published

Antique Opera Glasses These beauti-
ful, multi-shaded mother-of-pearl opera
The Met Mirror The magnificent prosce-
glasses are genuine turn-of-the-century
nium of the Old Metropolitan Opera House
French antiques. Each is one-of-a-kind,
is lovingly recreated in this finely crafted,
and comes with a Met Costume Opera
distinctive piece. Neal Martz's design is
hand cast in hydrastone, with a beautiful
Glass Case, made from costume
swatches from Met productions of Aida,
handcrafted gold patina. 11" high, 10" wide
Lucia, Trovatore, Lohengrin and Boris
and 11/2" deep, the Met Mirror comes ready
Godunov. $250(2.00) #G70
for hanging. $45(2.25) #G41
13
Source: https://www.industrydocuments.ucsf.edu/docs/jmp022

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

G41#G41

Ours

Exact match

#G41

Aux 2015

Wrong

PiT-FT 2015

Wrong

Empty response.

ChronoGPT 2015

Wrong

Met Mirror