Speculative decoding
VocabTrim
VOCABTRIM: Vocabulary Pruning for Efficient Speculative Decoding in LLMs
Superseded baseline#66 of 151 most-superseded · first seen Jun 28, 2025
Superseded — cited as a baseline and beaten by newer methods
1 papers critique it · 1 beat it on benchmarks
What papers say
Verbatim critique sentences, each from a paper that cites VocabTrim as a baseline.
All tokens outside this vocabulary are assigned zero probability and can never be proposed by the drafter, which typically reduces acceptance quality.
Beaten on benchmarks
Head-to-head results where a newer method reports beating VocabTrim. Values are copied from the source paper's tables — verify against the cited paper.
SlimSpec beats VocabTrim
2.40 vs 2.11
Speedup · [Llama3.1-8B-Instruct, temperature 1, batch size 1]
SlimSpec: Low-Rank Draft LM-Head for Accelerated Speculative Decoding
What to use instead
Recent methods in the same sub-problem, not yet superseded in the knowledge base — arXiv benchmark leaders, not vetted production recommendations.
- May 11, 2026
- EvoSpecEvoSpec: Evolving Speculative Decoding via Real-Time Vocabulary and Parameter AdaptationTargetApr 17, 2026
- Apr 8, 2026