Rethinking in Spikes: Mitigating Hallucinations in MDLMs with Step-Aware Decoding
cs.CL, cs.AI
Submitted: 2026-10-14
Updated: 2026-10-14
Project page: https://mdlm-steprefine.github.io
Terminology
Sources
- LLaDA2.0-Uni: Unifying Multimodal Understanding and Generation with Diffusion Large Language Model
- Learning to Parallel: Accelerating Diffusion Large Language Models via Learnable Parallel Decoding
- Accelerated Sampling from Masked Diffusion Models via Entropy Bounded Unmasking
- DFlash: Block Diffusion for Flash Speculative Decoding
- Unified Diffusion VLA: Vision-Language-Action Model via Joint Discrete Denoising Diffusion Process
- Beyond Surface Reasoning: Unveiling the True Long Chain-of-Thought Capacity of Diffusion Large Language Models
- DPad: Efficient Diffusion Language Models with Suffix Dropout
- dParallel: Learnable Parallel Decoding for dLLMs
- DMax: Aggressive Parallel Decoding for dLLMs
- Optimizing Decoding Paths in Masked Diffusion Models by Quantifying Uncertainty
- SDAR: A Synergistic Diffusion-AutoRegression Paradigm for Scalable Sequence Generation
- Improving Variable-Length Generation in Diffusion Language Models via Length Regularization
- Holistic Analysis of Hallucination in GPT-4V(ision): Bias and Interference Challenges
- Saber: An Efficient Sampling with Adaptive Acceleration and Backtracking Enhanced Remasking for Diffusion Language Model
- dVoting: Fast Voting for dLLMs
- From Bits to Rounds: Parallel Decoding with Exploration for Diffusion Language Models
- Self Speculative Decoding for Diffusion Large Language Models
- VidLaDA: Bidirectional Diffusion Large Language Models for Efficient Video Understanding
- Soft-Masked Diffusion Language Models
- Wide-In, Narrow-Out: Revokable Decoding for Efficient and Effective DLLMs
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