Large language models develop novel social biases through adaptive exploration

by paimapi | View on Hacker News