Dataset · Anthropic
Claude scorpion toxin protein design
Anthropic/scorpion-toxin-binder-design
Laboratory test results for proteins generated by the Claude model to bind scorpion toxins.
What's inside
A dataset from Anthropic containing results from a de novo protein design experiment using the Claude model to neutralize five scorpion toxins. Binding efficacy was tested in Adaptyv Bio and Twist Bioscience laboratories using surface plasmon resonance (SPR); out of all variants, 17 proteins proved functional. The dataset combines tabular data, text, and images (from 10K to 100K entries).
What you can do with it
- Analyzing and validating generative models for de novo protein design
- Training protein-protein binding prediction algorithms on real SPR measurements
- Investigating LLM capabilities in computational biology tasks
- Size
- 10K–100K rows
- Downloads this month
- 7K
- Likes
- 45
How to load it
from datasets import load_dataset
ds = load_dataset("Anthropic/scorpion-toxin-binder-design")