paper-with-me

홈 › Papers

Robust Reasoning via Dynamic Token Selection for Distribution-Aligned Self-Distillation

2026-05-30 · Ruiqi Zhang, Lingxiang Wang, Hainan Zhang Zhiming Zheng arxiv

Self-distillation improves learning efficiency by rewriting reference answers as training data that better matches the model's own distribution. However, reference answers also introduce strong stylistic biases, causing the generative model to imitate surface forms rather than learn useful reasoning patterns. We observe that the rewriting data contains a large number of high-perplexity (PPL) tokens, coming from two distinct sources: beneficial knowledge-enhancing logical corrections, and harmful stylistic drift induced by reference imitation. Treating all such tokens equally can disrupt the base model's original distribution and degrade performance, especially on difficult reasoning tasks. To address this, we propose Distribution-Aligned Self-Distillation (DASD), which uses an answer-aware reference model to generate candidate tokens and dynamically filters them according to the base model's confidence. DASD preserves tokens that encode useful logical knowledge while suppressing distributionally misaligned style noise. Experiments on math, code, and commonsense reasoning benchmarks show that DASD consistently outperforms competitive baselines, reduces high-PPL tokens, and improves robustness across tasks of varying difficulty.

📄 PDF Abstract BibTeX arXiv:2606.00628

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Entropy-Aware Branching for Improved Mathematical Reasoning

2025-03-27 · Xianzhi Li, Ethan Callanan, Xiaodan Zhu, Mathieu Sibue 외

While Large Language Models (LLMs) are effectively aligned through extensive pre-training and fine-tuning, they still struggle with varying levels of uncertainty during token generation. In our investigation of mathemati…

Mathematical Reasoning

Dynamic Thinking-Token Selection for Efficient Reasoning in Large Reasoning Models

2026-01-26 · Zhenyuan Guo, Tong Chen, Wenlong Meng, Chen Gong 외 arxiv

Large Reasoning Models (LRMs) excel at solving complex problems by explicitly generating a reasoning trace before deriving the final answer. However, these extended generations incur substantial memory footprint and comp…

Dynamic Collaboration of Multi-Language Models based on Minimal Complete Semantic Units

2025-08-26 · Chao Hao, Zezheng Wang, Yanhua Huang, Ruiwen Xu 외 arxiv

This paper investigates the enhancement of reasoning capabilities in language models through token-level multi-model collaboration. Our approach selects the optimal tokens from the next token distributions provided by mu…

Deeper is Not Always Better: Mitigating the Alignment Tax via Confident Layer Decoding

2026-06-20 · Xuanming Zhang, Sining Zhoubian, Yuxuan Chen, Tianyi Tang 외 arxiv

Autoregressive generation in large language models (LLMs) conventionally decodes from the final layer, assuming that deeper representations yield more reliable next-token predictions. We revisit this assumption by reveal…

PriFT: Prior-Support Guided Supervised Fine-Tuning

2026-06-08 · Ke Wang, Shuangqi Li, Mathieu Salzmann, Pascal Frossard arxiv

Supervised fine-tuning (SFT) is an efficient approach for downstream task adaptation and often serves as the initialization stage for reinforcement learning (RL), but it can show weaker generalization than RL. A key limi…

Reinforcement LearningMathematical ReasoningQuestion AnsweringCode Generation