Tag: efficient-llm

All the articles with the tag "efficient-llm".

• research

  15 Oct, 2026

  ProVe: Proposer-Verifier Decoding

  Speculative decoding with the roles reversed: the large model proposes a block of 16 tokens in one pass, a verifier reads it in one more. 2.5x exact (output equals greedy AR), 2.7x with a 3B verifier.

• research

  28 Jul, 2026

  MARS: Several Tokens per Forward Pass, Without Giving Up Autoregression

  One SFT checkpoint, two modes: full-quality AR at τ=1.0, or 1.8 tokens per forward pass for 1.56–2.16× speedup, with an exact, free fallback to plain AR.