Marc Finzi (@m_finzi) 's Twitter Profile
Marc Finzi

@m_finzi

Postdoc at CMU in Locus Lab -- Machine Learning

ID: 816696160618475520

linkhttps://mfinzi.github.io/ calendar_today04-01-2017 17:21:52

155 Tweet

958 Followers

308 Following

Marc Finzi (@m_finzi) 's Twitter Profile Photo

Why do larger language models generalize better? In our new ICLR paper, we derive an interpretable generalization bound showing that compute-optimal LLMs provably generalize better with scale! 📄arxiv.org/abs/2504.15208 1/7🧵