Home BlogDataset Evaluating Language Models

Evaluating Language Models

by Morgan Stevens

Anthropic, a U.S.-based AI company, has created a dataset to evaluate language models that are used in high-impact decisions, such as determining financing or housing eligibility. The dataset contains prompts that decision-makers may input into language models for 70 diverse decision types and identifies potential discriminatory impact, including both positive and negative discrimination. 

Get the data.

Image credit: Flickr user Ars Electronica

You may also like

Show Buttons
Hide Buttons