Get started
Author

Xiaosong Yuan

SG-OPD: Sign-Gated On-Policy Distillation via Sign-Consistency Gating and Phased Teacher Sampling
2026 · Xu, Haoran, Wang, Hongyu, et al. · arXiv (Cornell University) · 14
PDF
Backtracking When It Strays: Mitigating Dual Exposure Biases in LLM Reasoning Distillation
2026 · Wang, Bing, Yan, Shaotian, et al. · arXiv (Cornell University) · 13
PDF
Hallucination Begins Where Saliency Drops
2026 · Zhang, Xiaofeng, Zhu, Yuanchao, et al. · arXiv (Cornell University) · 7
PDF
Coalition-Aware Skill Reliability for Self-Evolving Agents
2026 · Zhao, Qiyan, Zhang, Xiaofeng, et al. · arXiv (Cornell University) · 6
PDF
Context Tokens are Anchors: Understanding the Repetition Curse in dMLLMs from an Information Flow Perspective
2026 · Zhao, Qiyan, Zhang, Xiaofeng, et al. · arXiv (Cornell University) · 3
PDF
Visual Para-Thinker++: A Single-Policy Multi-Agent Framework for Visual Reasoning
2026 · Xu, Haoran, Wang, Hongyu, et al. · arXiv (Cornell University) · 1
PDF
Are Rationales Necessary and Sufficient? Tuning LLMs for Explainable Misinformation Detection
2026 · Wang, Bing, Miao, Rui, et al. · arXiv (Cornell University) · 1
PDF
Context-Aware Cluster Decoding: Semantic Anchor-Driven Coherence in dMLLMs
2026 · Yikai Zhao, Qiyan Zhao, et al. · Conference on Empirical Methods in Natural Language Processing (EMNLP) · 1
PDF
ART: Attention Replacement Technique to Improve Factuality in LLMs
2026 · Association for Computational Linguistics 2026, Luo, Ziqin, et al. · Underline Science Inc.
Reasoning Fails Where Step Flow Breaks
2026 · Association for Computational Linguistics 2026, Pan, Yulan, et al. · Underline Science Inc.
Information-Guided Frontier Decoding: Contextual Utility-Driven Commitment in dMLLMs
2026 · Xingyou Fang, Jingxing Zhong, et al. · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP)
PDF
Works count: 11
Citation count: 46
H-index: 4
I10-index: 2
Year
Topic
Type