Allen Institute researchers find pervasive toxicity in popular language models

Researchers at the Allen Institute for AI find evidence of pervasive bias and toxicity in popular language models from OpenAI and Salesforce.Read More