STaRR: Spatial-Temporal Token-Dynamics-Aware Responsive Remasking for Diffusion Language Models
On the map: Diffusion models → Discrete and text diffusion
also in 1 more
Abstract (arXiv)
Diffusion Language Models (DLMs) enable parallel decoding via iterative denoising, where remasking strategies play a critical role in balancing inference speed and output quality. Existing methods predominantly rely on static confidence thresholds, overlooking the spatial-temporal dynamics of token confidence, causing unnecessary remasking. We propose Spatial-Temporal Token-Dynamics-Aware Responsive Remasking (STaRR), a training-free framework that dynamically adapts remasking decisions based on token confidence evolution. STaRR introduces two metrics, temporal variance and spatial deviance, to guide fine-grained, step-wise dynamic thresholding. We further introduce a step-wise dynamic thresholding strategy, further enhanced with responsiveness optimizations for scalability and robustness. Experiments show that STaRR achieves an average speedup of 4.1 and up to 8.9 while maintaining comparable accuracy.
Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Discrete and text diffusion, one of 14 directions in Diffusion models. Builds on 22 of its 23 references that are in the map. First posted on arXiv 7 Dec 2025.
Builds on
22 of 23 references.
Closest ancestors · 9
Same direction, newest first
- 2025Saber: An Efficient Sampling with Adaptive Acceleration and Backtracking Enhanced Remasking for Diffusion Language Model20 Oct
- 2025Dream 7B: Diffusion Large Language Models21 Aug
- 2025DPad: Efficient Diffusion Language Models with Suffix Dropout19 Aug
- 2025Time Is a Feature: Exploiting Temporal Dynamics in Diffusion Language Models12 Aug
- 2025Diffusion LLMs Can Do Faster-Than-AR Inference via Discrete Diffusion Forcing8 Aug
- 2025Plan for Speed: Dilated Scheduling for Masked Diffusion Language Models23 Jun
- 2025Discrete Diffusion in Large Language and Multimodal Models: A Survey16 Jun
- 2025dKV-Cache: The Cache for Diffusion Language Models21 May
- 2025Large Language Diffusion Models14 Feb
Other references · 13
- 2021Training Verifiers to Solve Math Word Problemscited by 9,743
- 2021Evaluating Large Language Models Trained on Codecited by 8,053
- 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,203
- 2021Program Synthesis with Large Language Modelscited by 3,509
- 2025Sparse-dLLM: Accelerating Diffusion LLMs with Dynamic Cache Evictioncited by 64
- 2025Wide-In, Narrow-Out: Revokable Decoding for Efficient and Effective DLLMscited by 45
- 2025Accelerating Diffusion Large Language Models with SlowFast Sampling: The Three Golden Principlescited by 43
- 2025AdaBlock-dLLM: Semantic-Aware Diffusion LLM Inference via Adaptive Block Sizecited by 30
- 2025Self Speculative Decoding for Diffusion Large Language Modelscited by 29
- 2025Thinking Inside the Mask: In-Place Prompting in Diffusion LLMscited by 13
- 2025Quant-dLLM: Post-Training Extreme Low-Bit Quantization for Diffusion Large Language Modelscited by 11
- 2025Rainbow Padding: Mitigating Early Termination in Instruction-Tuned Diffusion LLMscited by 10
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026NAVIRA: Decoupled Stochastic Remasking for Masked Diffusion Language Models
- 2026Re-evaluating Confidence Remasking in Masked Diffusion Language Models
- 2026CORE: Context-Robust Remasking for Diffusion Language Models
- 2026Remask, Don't Replace: Token-to-Mask Refinement in Diffusion Language Models
- 2026CoDR: Training-Free Confidence-Drift Remasking for Diffusion Language Models
- 2025Beyond Static Cutoffs: One-Shot Dynamic Thresholding for Diffusion Language Models
- 2025Don't Settle Too Early: Self-Reflective Remasking for Diffusion Language Models
- 2026CaRE Compute-aware Remasking Evaluation Protocol for Masked Diffusion Language Models