Skip to content
Read the original: Dwarkesh Patel· Published 18/100AI score18/100

Dwarkesh Patel on why Sonnet 5 and Opus 5 trail GLM 5.3

Original titleWas really interesting to hear John, Beren, and Charlie speculate about why Sonnet 5 and Opus 5 feel like worse models than GLM 5.3

AISummary

Dwarkesh Patel said a discussion with John, Beren, and Charlie questioned why Sonnet 5 and Opus 5 feel weaker than GLM 5.3, even though Anthropic could use raw logit distillation from Fable and train on Fable's environments. The discussion raised questions about the value of distillation, what makes it effective, and which model behaviors are hard to extract through it.

Read the original x.com

Source: Dwarkesh Patel · x.comPublished · added here