LLaDA (Speculative decoding): superseded — cited as a baseline and beaten by newer methods. 0 paper(s) critique it, 2 beat it on benchmarks — #46 of 151 most-superseded. Sub-problem: cluster led by SSD. Newer alternatives in the same sub-problem include FeF-DLLM, DualDiffusion, Principled Coarse-Graining (PCG).

Superseded baseline#46 of 151 most-superseded

LLaDA

Large Language Diffusion Models

Speculative decoding · first seen Feb 14, 2025

superseded — cited as a baseline and beaten by newer methods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 LLaDA. Values are copied from the source paper's tables — verify against the cited paper.

DualDiffusion beats LLaDA · Time (s) [no_condition]
82.0 vs 320.5
DualDiffusion: A Speculative Decoding Strategy for Masked Diffusion Models
FeF-DLLM (step=2) beats LLaDA · Acc. [GSM8K]
79.38 vs 78.60
Factorization-Error-Free Discrete Diffusion Language Model via Speculative Decoding
FeF-DLLM (step=2) beats LLaDA · Acc. [MATH]
36.40 vs 26.60
Factorization-Error-Free Discrete Diffusion Language Model via Speculative Decoding
FeF-DLLM (step=2) beats LLaDA · Acc. [HumanEval]
48.78 vs 47.60
Factorization-Error-Free Discrete Diffusion Language Model via Speculative Decoding
FeF-DLLM (step=2) beats LLaDA · Acc. [MBPP]
42.60 vs 34.20
Factorization-Error-Free Discrete Diffusion Language Model via Speculative Decoding
FeF-DLLM (step=2) beats LLaDA · Acc. [Mean]
51.79 vs 46.75
Factorization-Error-Free Discrete Diffusion Language Model via Speculative Decoding
FeF-DLLM (step=4) beats LLaDA · Acc. [GSM8K]
79.68 vs 78.60
Factorization-Error-Free Discrete Diffusion Language Model via Speculative Decoding
FeF-DLLM (step=4) beats LLaDA · Acc. [MATH]
36.56 vs 26.60
Factorization-Error-Free Discrete Diffusion Language Model via Speculative Decoding
FeF-DLLM (step=4) beats LLaDA · Acc. [HumanEval]
49.39 vs 47.60
Factorization-Error-Free Discrete Diffusion Language Model via Speculative Decoding
FeF-DLLM (step=4) beats LLaDA · Acc. [MBPP]
42.60 vs 34.20
Factorization-Error-Free Discrete Diffusion Language Model via Speculative Decoding
FeF-DLLM (step=4) beats LLaDA · Acc. [Mean]
52.06 vs 46.75
Factorization-Error-Free Discrete Diffusion Language Model via Speculative Decoding

Newer alternatives

Recent methods in the same sub-problem, not yet superseded in the knowledge base.