Image de couverture de livre

Blockwise SFT for Diffusion Language Models: Reconciling Bidirectional Attention and Autoregressive Decoding

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Sun, Bowen, Cai, Yujun, Yang, Ming-Hsuan, Wang, Yiwei
Format:	Preprint
Publié:	2025
Sujets:	Computation and Language
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Unveiling the Potential of Diffusion Large Language Model in Controllable Generation
par: Xiong, Zhen, et autres
Publié: (2025)

Structured Attention Matters to Multimodal LLMs in Document Understanding
par: Liu, Chang, et autres
Publié: (2025)

MRFD: Multi-Region Fusion Decoding with Self-Consistency for Mitigating Hallucinations in LVLMs
par: Ge, Haonan, et autres
Publié: (2025)

Causal Autoregressive Diffusion Language Model
par: Ruan, Junhao, et autres
Publié: (2026)

Symbolic or Numerical? Understanding Physics Problem Solving in Reasoning LLMs
par: Dan, Nifu, et autres
Publié: (2025)

PACER: Blockwise Pre-verification for Speculative Decoding with Adaptive Length
par: Zhang, Situo, et autres
Publié: (2026)

Exploring and Improving Drafts in Blockwise Parallel Decoding
par: Kim, Taehyeon, et autres
Publié: (2024)

Con-ReCall: Detecting Pre-training Data in LLMs via Contrastive Decoding
par: Wang, Cheng, et autres
Publié: (2024)

Finding Distributed Object-Centric Properties in Self-Supervised Transformers
par: Rawlekar, Samyak, et autres
Publié: (2026)

Do "New Snow Tablets" Contain Snow? Large Language Models Over-Rely on Names to Identify Ingredients of Chinese Drugs
par: Li, Sifan, et autres
Publié: (2025)

PAS: A Training-Free Stabilizer for Temporal Encoding in Video LLMs
par: Sun, Bowen, et autres
Publié: (2025)

Thinking with Sound: Audio Chain-of-Thought Enables Multimodal Reasoning in Large Audio-Language Models
par: Xiong, Zhen, et autres
Publié: (2025)

SemVink: Advancing VLMs' Semantic Understanding of Optical Illusions via Visual Global Thinking
par: Li, Sifan, et autres
Publié: (2025)

MIRAGE: Multimodal Immersive Reasoning and Guided Exploration for Red-Team Jailbreak Attacks
par: You, Wenhao, et autres
Publié: (2025)

WeDLM: Reconciling Diffusion Language Models with Standard Causal Attention for Fast Inference
par: Liu, Aiwei, et autres
Publié: (2025)

Why Diffusion Language Models Struggle with Truly Parallel (Non-Autoregressive) Decoding?
par: Li, Pengxiang, et autres
Publié: (2026)

Falcon: Faster and Parallel Inference of Large Language Models through Enhanced Semi-Autoregressive Drafting and Custom-Designed Decoding Tree
par: Gao, Xiangxiang, et autres
Publié: (2024)

Adaptive Blockwise Search: Inference-Time Alignment for Large Language Models
par: Quamar, Mohammad Atif, et autres
Publié: (2025)

Bitune: Leveraging Bidirectional Attention to Improve Decoder-Only LLMs
par: Kopiczko, Dawid J., et autres
Publié: (2024)

Texture or Semantics? Vision-Language Models Get Lost in Font Recognition
par: Li, Zhecheng, et autres
Publié: (2025)

Mapping the Minds of LLMs: A Graph-Based Analysis of Reasoning LLM
par: Xiong, Zhen, et autres
Publié: (2025)

GeoSVG-RL: Geometry-Aware Reinforcement Learning for Layout-Constrained Text-to-SVG Diagram Generation
par: Li, Sifan, et autres
Publié: (2026)

ReasonGen-R1: CoT for Autoregressive Image generation models through SFT and RL
par: Zhang, Yu, et autres
Publié: (2025)

Understanding GUI Agent Localization Biases through Logit Sharpness
par: Tao, Xingjian, et autres
Publié: (2025)

Are LLMs Really Not Knowledgeable? Mining the Submerged Knowledge in LLMs' Memory
par: Tao, Xingjian, et autres
Publié: (2024)

A Factuality and Diversity Reconciled Decoding Method for Knowledge-Grounded Dialogue Generation
par: Yang, Chenxu, et autres
Publié: (2024)

ReFusion: A Diffusion Large Language Model with Parallel Autoregressive Decoding
par: Li, Jia-Nan, et autres
Publié: (2025)

$A^2R^2$: Advancing Img2LaTeX Conversion via Visual Reasoning with Attention-Guided Refinement
par: Li, Zhecheng, et autres
Publié: (2025)

FourierSampler: Unlocking Non-Autoregressive Potential in Diffusion Language Models via Frequency-Guided Generation
par: He, Siyang, et autres
Publié: (2026)

Bidirectional Awareness Induction in Autoregressive Seq2Seq Models
par: Hu, Jia Cheng, et autres
Publié: (2024)

An Empirical Study of SFT-DPO Interaction and Parameterization in Small Language Models
par: Feng, Yuming, et autres
Publié: (2026)

When Linear Attention Meets Autoregressive Decoding: Towards More Effective and Efficient Linearized Large Language Models
par: You, Haoran, et autres
Publié: (2024)

Vision Language Models Map Logos to Text via Semantic Entanglement in the Visual Projector
par: Li, Sifan, et autres
Publié: (2025)

ASCD: Attention-Steerable Contrastive Decoding for Reducing Hallucination in MLLM
par: Wang, Yujun, et autres
Publié: (2025)

SparseD: Sparse Attention for Diffusion Language Models
par: Wang, Zeqing, et autres
Publié: (2025)

AR-MAP: Are Autoregressive Large Language Models Implicit Teachers for Diffusion Large Language Models?
par: Lin, Liang, et autres
Publié: (2026)

DiffuSpec: Unlocking Diffusion Language Models for Speculative Decoding
par: Li, Guanghao, et autres
Publié: (2025)

Reversible Diffusion Decoding for Diffusion Language Models
par: Wang, Xinyun, et autres
Publié: (2026)

Scaling Diffusion Language Models via Adaptation from Autoregressive Models
par: Gong, Shansan, et autres
Publié: (2024)

BESA: Pruning Large Language Models with Blockwise Parameter-Efficient Sparsity Allocation
par: Xu, Peng, et autres
Publié: (2024)