11 papers · 1 filter
Negative Self-Distillation: Learning to Reason by Avoiding Flaws
Rongcan Pei, Zhepei Wei, Shuyao Xu +3
On-Policy Self-Distillation (OPSD) has emerged as a popular paradigm for large language model (LLM) self-improvement, allowing models to act as their own teachers by leveraging pri…
Separate the Wheat from the Chaff: Winnowing Down Divergent Views in Retrieval Augmented Generation
Song Wang, Zihan Chen, Peng Wang +5
Retrieval-augmented generation (RAG) enhances large language models (LLMs) by integrating external knowledge sources to address their limitations in accessing up-to-date or special…
Aligning Large Language Models via Fully Self-Synthetic Data
Shangjian Yin, Zhepei Wei, Xinyu Zhu +2
Traditional reinforcement learning from human feedback (RLHF) for large language models (LLMs) relies on expensive human-annotated datasets, while Reinforcement Learning from AI Fe…
TruthRL: Incentivizing Truthful LLMs via Reinforcement Learning
Zhepei Wei, Xiao Yang, Kai Sun +12
While large language models (LLMs) have demonstrated strong performance on factoid question answering, they are still prone to hallucination and untruthful responses, particularly…
AdaDecode: Accelerating LLM Decoding with Adaptive Layer Parallelism
Zhepei Wei, Wei-Lin Chen, Xinyu Zhu +1
Large language models (LLMs) are increasingly used for long-content generation (e.g., long Chain-of-Thought reasoning) where decoding efficiency becomes a critical bottleneck: Auto…
The Surprising Effectiveness of Negative Reinforcement in LLM Reasoning
Xinyu Zhu, Mengzhou Xia, Zhepei Wei +3
Reinforcement learning with verifiable rewards (RLVR) is a promising approach for training language models (LMs) on reasoning tasks that elicit emergent long chains of thought (CoT…