Is AdaEDL superseded?

AdaEDL (Speculative decoding): superseded — cited as a baseline and beaten by newer methods. 0 paper(s) critique it, 2 beat it on benchmarks — #43 of 151 most-superseded. Sub-problem: cluster led by SpecInfer. Newer alternatives in the same sub-problem include SpecKV, FASER, ConfLayers, Goose, greedy multi-path block verification (GBV).

Speculative decoding

AdaEDL

AdaEDL: Early Draft Stopping for Speculative Decoding of Large Language Models via an Entropy-based Lower Bound on Token Acceptance Probability

Superseded baseline#43 of 151 most-superseded · first seen Oct 24, 2024

Superseded — cited as a baseline and beaten by newer methods

0 papers critique it · 2 beat it on benchmarks

Beaten on benchmarks

Head-to-head results where a newer method reports beating AdaEDL. Values are copied from the source paper's tables — verify against the cited paper.

WVIR-based beats AdaEDL
198 vs 284
Percentile increment in mean latency · [WMT14 dataset, low-acceptance-rate regime (Gemma 27B/2B)]
DSDE: Dynamic Speculative Decoding with KLD Stability for Real-World Serving
Pacer beats AdaEDL
17.95 vs 17.47
tokens/s (Average) · [SpecBench, Llama-2 7B/70B]
PACER: Blockwise Pre-verification for Speculative Decoding with Adaptive Length

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Recent methods in the same sub-problem, not yet superseded in the knowledge base — arXiv benchmark leaders, not vetted production recommendations.