Back to papers
August 4, 2026cs.CLcs.AIcs.LG
Efficient Knowledge Distillation for LLMs: Offline Top-K Logits and a Fused Chunked KL Loss
Bakbergen Ryskulov, Iker García-Ferrero, David Montero, David Jansen, Ali Hashemi, Jezabel R. Garcia, Antonio Tiene, Román Orús
HF Upvotes
9
Categories
cs.CL, cs.AI, cs.LG