Cover Image

$ρ$-$\texttt{EOS}$: Training-free Bidirectional Variable-Length Control for Masked Diffusion LLMs

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Yang, Jingyi, Jiang, Yuxian, Shao, Jing
Format:	Preprint
Published:	2026
Subjects:	Computation and Language
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Diffusion LLM with Native Variable Generation Lengths: Let [EOS] Lead the Way
by: Yang, Yicun, et al.
Published: (2025)

DARE: Diffusion Large Language Models Alignment and Reinforcement Executor
by: Yang, Jingyi, et al.
Published: (2026)

Controlling Summarization Length Through EOS Token Weighting
by: Belligoli, Zeno, et al.
Published: (2025)

Taming Masked Diffusion Language Models via Consistency Trajectory Reinforcement Learning with Fewer Decoding Step
by: Yang, Jingyi, et al.
Published: (2025)

The Devil is in the EOS: Sequence Training for Detailed Image Captioning
by: Mohamed, Abdelrahman, et al.
Published: (2025)

Beyond Fixed: Training-Free Variable-Length Denoising for Diffusion Large Language Models
by: Li, Jinsong, et al.
Published: (2025)

\texttt{ReMind}: Understanding Deductive Code Reasoning in LLMs
by: Gao, Jun, et al.
Published: (2025)

Improving Variable-Length Generation in Diffusion Language Models via Length Regularization
by: Cheng, Zicong, et al.
Published: (2026)

Conditional Advantage Estimation for Reinforcement Learning in Large Reasoning Models
by: Chen, Guanxu, et al.
Published: (2025)

Evolutionary Guided Decoding: Iterative Value Refinement for LLMs
by: Liu, Zhenhua, et al.
Published: (2025)

Length Controlled Generation for Black-box LLMs
by: Gu, Yuxuan, et al.
Published: (2024)

Thinking Inside the Mask: In-Place Prompting in Diffusion LLMs
by: Jin, Xiangqi, et al.
Published: (2025)

Training LLMs to be Better Text Embedders through Bidirectional Reconstruction
by: Su, Chang, et al.
Published: (2025)

Masked Diffusion Language Models with Frequency-Informed Training
by: Kosmopoulou, Despoina, et al.
Published: (2025)

Prompting LLMs: Length Control for Isometric Machine Translation
by: Javorský, Dávid, et al.
Published: (2025)

$\texttt{SEM-CTRL}$: Semantically Controlled Decoding
by: Albinhassan, Mohammad, et al.
Published: (2025)

Dataset Decomposition: Faster LLM Training with Variable Sequence Length Curriculum
by: Pouransari, Hadi, et al.
Published: (2024)

$\texttt{BluePrint}$: A Social Media User Dataset for LLM Persona Evaluation and Training
by: Bück-Kaeffer, Aurélien, et al.
Published: (2025)

Rethinking Entropy Regularization in Large Reasoning Models
by: Jiang, Yuxian, et al.
Published: (2025)

$\texttt{LM}^\texttt{2}$: A Simple Society of Language Models Solves Complex Reasoning
by: Juneja, Gurusha, et al.
Published: (2024)

MiniPLM: Knowledge Distillation for Pre-Training Language Models
by: Gu, Yuxian, et al.
Published: (2024)

Less is More: Mitigating Multimodal Hallucination from an EOS Decision Perspective
by: Yue, Zihao, et al.
Published: (2024)

Training-free Truthfulness Detection via Value Vectors in LLMs
by: Liu, Runheng, et al.
Published: (2025)

Reconsidering Positional Supervision in Masked Diffusion Language Model Training
by: Ye, Mengyu, et al.
Published: (2026)

Intermittent Semi-Working Mask: A New Masking Paradigm for LLMs
by: Hu, HaoYuan, et al.
Published: (2024)

Autoregressive vs. Masked Diffusion Language Models: A Controlled Comparison
by: Vicentino, Caio
Published: (2026)

Diffusion-Inspired Masked Fine-Tuning for Knowledge Injection in Autoregressive LLMs
by: Pan, Xu, et al.
Published: (2025)

Understanding and Accelerating the Training of Masked Diffusion Language Models
by: Hong, Chunsan, et al.
Published: (2026)

KV-Embedding: Training-free Text Embedding via Internal KV Re-routing in Decoder-only LLMs
by: Tang, Yixuan, et al.
Published: (2026)

Instruction Pre-Training: Language Models are Supervised Multitask Learners
by: Cheng, Daixuan, et al.
Published: (2024)

Prompt-Based One-Shot Exact Length-Controlled Generation with LLMs
by: Xie, Juncheng, et al.
Published: (2025)

Length-Unbiased Sequence Policy Optimization: Revealing and Controlling Response Length Variation in RLVR
by: Liu, Fanfan, et al.
Published: (2026)

Blockwise SFT for Diffusion Language Models: Reconciling Bidirectional Attention and Autoregressive Decoding
by: Sun, Bowen, et al.
Published: (2025)

Improving Sampling for Masked Diffusion Models via Information Gain
by: Yang, Kaisen, et al.
Published: (2026)

MultiTok: Variable-Length Tokenization for Efficient LLMs Adapted from LZW Compression
by: Elias, Noel, et al.
Published: (2024)

Revise, Don't Freeze: Sampler-Matched Training for Self-Correcting Masked Diffusion Language Models
by: Yu, Longxuan, et al.
Published: (2026)

Bitune: Leveraging Bidirectional Attention to Improve Decoder-Only LLMs
by: Kopiczko, Dawid J., et al.
Published: (2024)

Reward-Weighted Sampling: Enhancing Non-Autoregressive Characteristics in Masked Diffusion LLMs
by: Gwak, Daehoon, et al.
Published: (2025)

Beyond Single-Task: Robust Multi-Task Length Generalization for LLMs
by: Hu, Yi, et al.
Published: (2025)

Plan-and-Write: Structure-Guided Length Control for LLMs without Model Retraining
by: Akinfaderin, Adewale, et al.
Published: (2025)