On this day: Nov 21

Lets servers provide interactive UIs via the ui:// scheme and sandboxed iframes; co-authored by competing camps

Named the approach of replacing the reward model with a verifier function, and released every stage of the pipeline

The context-length race entered the 200,000-token range

A community test settled in as the de facto standard evaluation, establishing the recognition that the stated context length and actual recall performance differ

On 466 questions, humans scored 92% versus 15% for GPT-4 with plugins; shifted the focus of evaluation from hard exams to practical ability

A redesign of FSDP1. The precondition that made FP8 training, partial freezing and compiler optimization possible