Skip to content

Dataset · futurehouse

Biological research benchmark

futurehouse/lab-bench

A set of text and visual tasks to evaluate AI agent capabilities in scientific biology.

What's inside

It contains questions and answers in text and image formats (ranging from 1,000 to 10,000 examples), divided into 8 categories and 30 subtasks. The data covers fact extraction from scientific papers, database search, supplementary material analysis, and scientific graph interpretation.

What you can do with it

  • Evaluating scientific AI agents in biology
  • Testing multimodal models on understanding scientific illustrations and graphs
  • Checking scientific literature and database information retrieval systems
Tasks
Question answering
Size
1K–10K rows
Downloads this month
21.3K
Likes
54

How to load it

from datasets import load_dataset

ds = load_dataset("futurehouse/lab-bench")