Hacker News new | past | comments | ask | show | jobs | submit | from login

Calculating the cost of a Google DeepMind paper (152334h.github.io) 303 points by 152334H 6 months ago | past | 150 comments Knowing Enough About MoE to Explain Dropped Tokens in GPT-4 (152334h.github.io) 3 points by 152334H on Aug 8, 2023 | past | 1 comment Non-determinism in GPT-4 is caused by Sparse MoE (152334h.github.io) 397 points by 152334H on Aug 4, 2023 | past | 181 comments Why can't TorToiSe be fine-tuned? (152334h.github.io) 1 point by 152334H on Feb 11, 2023 | past

Consider applying for YC's Spring batch! Applications are open till Feb 11. Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact Search: