AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

WeDLM: Reconciling Diffusion Language Models with Standard Causal Attention for Fast Inference

Aiwei Liu, Minghua He, Shaoxun Zeng, Sijun Zhang, Linhao Zhang, Chuhan Wu, Wei Jia, Yuan Liu, Zhou Xiao, Jie Zhou

ICML 2026 · Large language models › Language, safety and interpretability · approach: Parallel and early-exit decoding · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace block-wise decoding with streaming parallel decoding (ablation)WeDLM-7B-Instruct · MATH, tau=0.9tokens/s: 221 -> 423 tokens/s (1.9x)Improved
increase distance penalty lambda from 0.01 to 0.05 (ablation)WeDLM-7B-Instruct · MATHaccuracy: +2.6 points with 3% speed reductionImproved
raise entropy threshold tau above 0.5 (ablation)WeDLM-7B-Instruct · MATHaccuracyHurt
use causal intra-block attention instead of bidirectional intra-block attention (ablation)WeDLM · average over benchmarks (AR-initialized)average scoreImproved
use diffusion-style training instead of continued AR pretraining (7B) (ablation)WeDLM-7B · average over benchmarksaverage score: +3.5 points over AR continueImproved
apply WeDLM adaptation to small models (0.6B, 1.5B) (ablation)WeDLM · average over benchmarksaverage scoreHurt
vary training block size B among 4, 8, 32 (ablation)WeDLM-7B · average over benchmarksaverage score: 71.23 / 70.45 / 70.84No clear effect
convert Qwen3-8B-Instruct into WeDLM-8B-InstructWeDLM-8B · average over 9 benchmarksaverage score: 75.12 -> 77.53Improved
convert Qwen3-8B-Instruct into WeDLM-8B-InstructWeDLM-8B · MATH (4-shot)accuracy: 69.60 -> 64.80Hurt
convert Qwen3-8B-Instruct into WeDLM-8B-InstructWeDLM-8B · HumanEval (4-shot)accuracy: 71.95 -> 80.49Improved
convert base models to WeDLMWeDLM-7B/8B base · MBPPaccuracy: 3-4 points dropHurt
raise tau from 0.2 to 0.9 in streaming decoding (ablation)WeDLM-8B-Instruct · GSM8Kaccuracy / speedup: 92.3% at 1.97x -> above 79% at 3.2xMixed

Later papers that compared with this method

1 later accepted paper reports WeDLM-8B as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy