Community Paper Reading: Why Language Models Hallucinate

​Join our upcoming community paper reading, where we'll dive into the paper published by OpenAI: "Why Language Models Hallucinate."

We're thrilled to host one of the paper's authors — Santosh S. Vempala, Frederick Storey II Chair of Computing and Distinguished Professor in the School of Computer Science, Georgia Tech — who will walk us through the research and its implications. If time permits, there will be a live Q&A session, so bring your questions!

The paper argues that hallucinations persist due to the way most evaluations are graded—language models are optimized to be good test-takers, and guessing when uncertain improves test performance. This “epidemic” of penalizing uncertain responses can only be addressed through a socio-technical mitigation: modifying the scoring of existing benchmarks that are misaligned but dominate leaderboards, rather than introducing additional hallucination evaluations.

Read the paper: https://cdn.openai.com/pdf/d04913be-3f6f-4d2b-b283-ff432ef4aaa5/why-language-models-hallucinate.pdf