Concept
Acceptance length
Acceptance length is the number of draft tokens accepted per verification step in speculative decoding. In speculative decoding, a cheap draft model (or a diffusion block) proposes several tokens at once, and the expensive target model verifies them in a single forward pass.
The rest of “Acceptance length” is a premium feature: every concept in the library gets a precise, practitioner-focused write-up like this one, cross-linked straight from the paper summaries that use it.
Log in to unlock→