paper-with-me

Papers

Learning from the Self-future: On-policy Self-distillation for dLLMs

2026-06-16 · Yifu Luo, Zeyu Chen, Haoyu Wang, Xinhao Hu, Yuxuan Zhang, Zhizhou Sha, Shiwei Liu arxiv

On-policy self-distillation (OPSD) has proven effective for post-training large language models (LLMs), yet its application to diffusion LLMs (dLLMs) remains unexplored. Existing OPSD methods are inherently autoregressive-centric. They inject privileged information via left-to-right prefix conditioning with token-level divergence supervision, a design that fundamentally conflicts with the arbitraryorder generation of dLLMs. We introduce d-OPSD, the first OPSD framework tailored for dLLMs. Our approach makes two core contributions. First, we reframe self-teacher construction by using self-generated answers as suffix conditioning, enabling the student model to learn from "self future-experience" rather than privileged prefixes. Second, we shift supervision from token-level to step-level, aligning training with the iterative denoising process of dLLMs. Experiments across four reasoning benchmarks show that d-OPSD consistently outperforms RLVR and SFT baselines with superior sample efficiency, requiring only around 10% of the optimization steps by RLVR and opening a promising pathway for dLLM posttraining. The code is available at https://github.com/xingzhejun/d-OPSD.

📄 PDF Abstract BibTeX arXiv:2606.18195

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

GDSD: Reinforcement Learning as Guided Denoiser Self-Distillation for Diffusion Language Models

2026-05-28 · Xiaohang Tang, Keyue Jiang, Che Liu, Qifang Zhao 외 arxiv

Reinforcement learning (RL) can be used to improve the policy (denoiser) of diffusion large language models (dLLMs), while being hindered by the intractability of the policy likelihood. A dominant and efficient family of…

Reinforcement Learning

dOPSD: On-Policy Self-Distillation for Diffusion Language Models

2026-07-05 · Phuong Tuan Dat, Qi Li, Xinchao Wang hf

Diffusion large language models (dLLMs) generate text by iteratively denoising a masked sequence, offering a parallel alternative to autoregressive models, but eliciting strong reasoning through post-training remains dif…

Reinforcement LearningCode Generation

When to Commit? Towards Variable-Size Self-Contained Blocks for Discrete Diffusion Language Models

2026-04-27 · Danny Wang, Ruihong Qiu, Zi Huang arxiv

Discrete diffusion language models (dLLMs) enable parallel token updates with bidirectional attention, yet practical generation typically adopts blockwise semi-autoregressive decoding. This switch creates a training-infe…

CForce: Boosting Parallel Decoding for dLLMs via Consistency Forcing

2026-08-14 · Yuji Ren, Chenkai Xu, Zhuocheng Gong, Jianguo Li 외 arxiv

Diffusion large language models (dLLMs) accelerate language generation by predicting multiple masks in a single forward pass. However, existing dLLMs can suffer from unreliable predictions in early denoising stages under…

Stabilizing Reinforcement Learning for Diffusion Language Models

2026-03-06 · Jianyuan Zhong, Kaibo Wang, Ding Ding, Zijin Feng 외 arxiv

Group Relative Policy Optimization (GRPO) is highly effective for post-training autoregressive (AR) language models, yet its direct application to diffusion large language models (dLLMs) often triggers reward collapse. W…

Reinforcement Learning