4 citations · 44 across the 49 of their papers we have counts for
13 papers · 1 filter
Reason-KE++: Aligning the Process, Not Just the Outcome, for Faithful LLM Knowledge Editing
Yuchen Wu, Liang Ding, Li Shen +1
Aligning Large Language Models (LLMs) to be faithful to new knowledge in complex, multi-hop reasoning tasks is a critical, yet unsolved, challenge. We find that SFT-based methods,…
Information-Theoretic Reward Modeling for Stable RLHF: Detecting and Mitigating Reward Hacking
Yuchun Miao, Liang Ding, Sen Zhang +3
Despite the success of Reinforcement Learning from Human Feedback (RLHF) in aligning language models with human values, reward hacking-or reward over-optimization-remains a major c…
Robust Knowledge Editing via Explicit Reasoning Chains for Distractor-Resilient Multi-Hop QA
Yuchen Wu, Liang Ding, Li Shen +1
Large language models (LLMs) encode vast amounts of world knowledge but remain static once trained, making the timely integration of emerging facts prohibitively expensive via full…
Revisiting Overthinking in Long Chain-of-Thought from the Perspective of Self-Doubt
Keqin Peng, Liang Ding, Yuanxin Ouyang +2
Reasoning Large Language Models (RLLMs) have demonstrated impressive performance on complex tasks, largely due to the adoption of Long Chain-of-Thought (Long CoT) reasoning. Howeve…
KaFT: Knowledge-aware Fine-tuning for Boosting LLMs' Domain-specific Question-Answering Performance
Qihuang Zhong, Liang Ding, Xiantao Cai +3
Supervised fine-tuning (SFT) is a common approach to improve the domain-specific question-answering (QA) performance of large language models (LLMs). However, recent literature rev…
Runaway is Ashamed, But Helpful: On the Early-Exit Behavior of Large Language Model-based Agents in Embodied Environments
Qingyu Lu, Liang Ding, Siyi Cao +4
Agents powered by large language models (LLMs) have demonstrated strong planning and decision-making capabilities in complex embodied environments. However, such agents often suffe…
Entropy-Guided Watermarking for LLMs: A Test-Time Framework for Robust and Traceable Text Generation
Shizhan Cai, Liang Ding, Dacheng Tao
The rapid development of Large Language Models (LLMs) has intensified concerns about content traceability and potential misuse. Existing watermarking schemes for sampled text often…
VisuoThink: Empowering LVLM Reasoning with Multimodal Tree Search
Yikun Wang, Siyin Wang, Qinyuan Cheng +5
Recent advancements in Large Vision-Language Models have showcased remarkable capabilities. However, they often falter when confronted with complex reasoning tasks that humans typi…
Retrieval-Augmented Perception: High-Resolution Image Perception Meets Visual RAG
Wenbin Wang, Yongcheng Jing, Liang Ding +5
High-resolution (HR) image perception remains a key challenge in multimodal large language models (MLLMs). To overcome the limitations of existing methods, this paper shifts away f…
Edit Once, Update Everywhere: A Simple Framework for Cross-Lingual Knowledge Synchronization in LLMs
Yuchen Wu, Liang Ding, Li Shen +1
Knowledge editing allows for efficient adaptation of large language models (LLMs) to new information or corrections without requiring full retraining. However, prior methods typica…
Enhancing Input-Label Mapping in In-Context Learning with Contrastive Decoding
Keqin Peng, Liang Ding, Yuanxin Ouyang +3
Large language models (LLMs) excel at a range of tasks through in-context learning (ICL), where only a few task examples guide their predictions. However, prior research highlights…
TeZO: Empowering the Low-Rankness on the Temporal Dimension in the Zeroth-Order Optimization for Fine-tuning LLMs
Yan Sun, Tiansheng Huang, Liang Ding +2
Zeroth-order optimization (ZO) has demonstrated remarkable promise in efficient fine-tuning tasks for Large Language Models (LLMs). In particular, recent advances incorporate the l…
The Energy Loss Phenomenon in RLHF: A New Perspective on Mitigating Reward Hacking
Yuchun Miao, Sen Zhang, Liang Ding +3
This work identifies the Energy Loss Phenomenon in Reinforcement Learning from Human Feedback (RLHF) and its connection to reward hacking. Specifically, energy loss in the final la…