DAWN: Dependency-Aware Fast Inference for Diffusion LLMs
cs.CL
Submitted: 2026-02-06
Updated: 2026-11-07
Comments: Accepted at NeurIPS 2026 (Main Track)
Code: https://github.com/lizhuo-luo/DAWN
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Structuring The Future: Diffusion LLM Speculative Decoding via Calibrated Draft Graphs
- Program Synthesis with Large Language Models
- Accelerated Sampling from Masked Diffusion Models via Entropy Bounded Unmasking
- The Reversal Curse: LLMs trained on "A is B" fail to learn "B is A"
- Stable Video Diffusion: Scaling Latent Video Diffusion Models to Large Datasets
- Evaluating Large Language Models Trained on Code
- dParallel: Learnable Parallel Decoding for dLLMs
- Training Verifiers to Solve Math Word Problems
- Measuring Mathematical Problem Solving With the MATH Dataset
- Wide-In, Narrow-Out: Revokable Decoding for Efficient and Effective DLLMs
- Accelerating Diffusion LLM Inference via Local Determinism Propagation
- HunyuanVideo: A Systematic Framework For Large Video Generative Models
- Mercury: Ultra-Fast Language Models Based on Diffusion
- Beyond Fixed: Training-Free Variable-Length Denoising for Diffusion Large Language Models
- WeDLM: Reconciling Diffusion Language Models with Standard Causal Attention for Fast Inference
- dLLM-Cache: Accelerating Diffusion Large Language Models with Adaptive Caching
- DSB: Dynamic Sliding Block Scheduling for Diffusion LLMs
- dKV-Cache: The Cache for Diffusion Language Models
- Large Language Diffusion Models
- GPT-4 Technical Report
Related papers
- ALARM: Audio-Language Alignment for Reasoning Models
- ThinkSum: Probabilistic reasoning over sets using large language models
- Coherence boosting: When your pretrained language model is not paying enough attention
- From Facts to Insights: A Persona-Driven Dual Memory Framework and Dataset for Role-Playing Agents
- Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization
- The Art of Asking: Multilingual Prompt Optimization for Synthetic Data