Tag: reverse-kl
All the articles with the tag "reverse-kl".
-
Notes on: MiniLLM
The reverse-KL policy-gradient derivation behind MiniLLM — per-token rewards, teacher-mixed sampling, length normalization — plus the actual GPT-2/OPT/LLaMA numbers.
All the articles with the tag "reverse-kl".
The reverse-KL policy-gradient derivation behind MiniLLM — per-token rewards, teacher-mixed sampling, length normalization — plus the actual GPT-2/OPT/LLaMA numbers.