Browse Teach Curate

Reasoning Models and Test-Time Compute

o1/DeepSeek-R1-style models: chain-of-thought as a learned behavior, reinforcement learning with verifiable rewards, and trading inference compute for accuracy.

Demerzel Video

Reasoning Models and Test-Time Compute

o1/DeepSeek-R1-style models: chain-of-thought as a learned behavior, reinforcement learning with verifiable rewards, and trading inference compute for accuracy.