causal

Soft labels carry far more information than hard labels

Distillation is so effective per sample because soft targets (the full logit distribution) have far higher entropy than a one-hot token answer, carrying much more information per sample; AlphaGo exploits this by training the policy network to imitate the full MCTS visit-count distribution rather than just the single selected action ('dark knowledge' distillation).

causalpending

Speaker

Eric Jang

Evidence Quote

“In AlphaGo, you don't train the policy network to imitate the MCTS action. You train it to imitate the MCTS distribution.”

Source

Dwarkesh Patel Podcast - Eric Jang – Building AlphaGo from scratch— Dwarkesh Patel Podcast
Created: 6/13/2026, 3:37:56 AM

Your notes

Loading notes...