claim
Large language models learn to produce text in the statistical style of their training corpus, including any errors present in that corpus.

Authors

Sources

Referenced by nodes (1)