~ similar to 2605.28456· 16 results
Longxuan Yu, Yunshu Wu, Yu Fu, Siheng Xiong +4 more
The paper introduces DSL-LLaDA, a method that lightly adapts a pre-trained masked diffusion language model to perform continuous denoising in embedding space, significantly improving text generation q…
Junxia Cui, Haotian Ye, Runchu Tian, Hongcan Guo +8 more
The paper proposes SimSD, a plug-and-play speculative decoding algorithm that adapts diffusion language models (dLLMs) to achieve fast, token-level acceleration by restoring causal masking capabilitie…
Xiaoyou Wu, Cheng-Jhih Shih, Binfei Ji, Yong Liu +1 more
BlockBatch introduces a novel framework that efficiently accelerates diffusion language model (dLLM) inference by simultaneously executing multiple block-size branches for a single request, achieving…
Zekai Li, Ji Liu, Yiqing Huang, Ziqiong Liu +2 more
The paper proposes a novel trace-aware decoding framework, combining Temporal-Spatial Parallel Decoding (TSPD) and Confidence Extrapolation (CE), to significantly accelerate the inference of diffusion…
The paper introduces NaRA, a noise-aware LoRA technique that dynamically adapts fine-tuning parameters based on the noise level during diffusion, significantly improving the performance of Diffusion L…
Sicheng Feng, Zigeng Chen, Gongfan Fang, Xinyin Ma +1 more
dMoE proposes a block-level Mixture-of-Experts (MoE) framework for Diffusion Large Language Models (dLLMs) that aggregates token-level expert distributions into a unified block-level distribution, sig…
Yuchen Zhu, Jing Shi, Chongjian Ge, Hao Tan +8 more
FLARE is a systematic conversion framework that enables a single checkpoint to support both autoregressive (AR) and diffusion-style parallel decoding for hybrid-attention large language models, achiev…
The paper introduces DLM-SWAI, a training-free method that effectively steers diffusion language models (DLMs) toward desired textual styles or properties by biasing the token distribution at each den…
Longxuan Yu, Shaorong Zhang, Yu Fu, Hui Liu +2 more
The paper introduces D3IM, a novel parameter-free sampler that enables direct revision of visible tokens in Masked Diffusion Language Models, and proposes SCOPE to mitigate the model's tendency to per…
The paper introduces MLLM-Microscope, a system that analyzes the internal structure of multimodal large language models (MLLMs), finding that modality fusion significantly impacts the linearity and di…
Sicheng Yang, Shulan Ruan, Shiwei Wu, Yu Liu +3 more
PolySpeech-100 introduces a massive, multi-lingual benchmark covering 110 linguistic variants to rigorously test Speech-LLMs, demonstrating that open-source models struggle with low-resource languages…
The paper analyzes order-agnostic language models (OALMs), finding that their learned conditionals are not true factorizations and proposing a variance-based diagnostic to compare the quality of diffe…
This paper analyzes the decoding process of masked diffusion models for graph-to-text generation, finding that structural fine-tuning disrupts natural entity-first generation and proposing a structura…
The paper introduces SHADOWMASK, the first systematic backdoor attack targeting Masked Diffusion Language Models (MDLMs), demonstrating near-100% attack success while preserving clean model utility.
The paper argues that using confidence-based decoding, which is optimized via training mask alignment, fundamentally misaligns Masked Diffusion Models (MDMs) from the logical flow needed for complex r…
VideoMLA introduces a novel Multi-Head Latent Attention (MLA) mechanism that replaces per-head KV caches with a shared low-rank content latent, significantly reducing memory and improving throughput f…