Skip to content

Dataset · m-a-p

SuperGPQA: LLM evaluation across 285 disciplines

m-a-p/SuperGPQA

English SuperGPQA benchmark for evaluating LLMs across 285 graduate disciplines.

What's inside

SuperGPQA is a text benchmark from the paper SuperGPQA: Scaling LLM Evaluation across 285 Graduate Disciplines. Data is in English, with a size of 10K to 100K examples. ODC-BY license; part of the content originates from other sources. The official leaderboard is promised later.

What you can do with it

  • Evaluate LLMs on graduate-level questions
  • Compare models across 285 disciplines
  • Test knowledge in narrow academic fields
Languages
English
Size
10K–100K rows
Downloads this month
21.3K
Likes
93

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("m-a-p/SuperGPQA")