14 papers
MemTools: A Unified Research Framework for Interoperable Agent Memory
Chengfeng Zhao, Jinhui Chen, Sirui Liang +4
While memory systems are essential for agent architectures, pervasive architectural fragmentation restricts systematic research. Existing implementations typically couple different…
JANUS: Foreseeing Latent Risk for Long-Horizon Agent Safety
Yuan Xiong, Linji Hao, Shizhu He +3
Agent safety is moving from content moderation toward preventing operational failures before tool-using agents act. We propose Janus, a foresight-oriented framework for long-horizo…
Break Through the Compression Bottleneck: From Theory to Practice
Xiusheng Huang, Lu Wang, Yequan Wang +2
As the parameter size of language models continues to grow, effective model compression is required to reduce their computational and memory overhead. Existing compression methods…
Neural Procedural Memory: Empowering LLM Agents with Implicit Activation Steering
Chengfeng Zhao, Yuqiao Tan, Shizhu He +3
While Large Language Models (LLMs) excel as static solvers, transforming them into autonomous agents remains challenging. This transition requires continuous environmental interact…
SkillPyramid: A Hierarchical Skill Consolidation Framework for Self-Evolving Agents
Yuan Xiong, Ziqi Miao, Qian Chen +6
Recent AI agents can flexibly invoke skills to solve complex tasks, but their long-term improvement is fundamentally constrained by a lack of systematic skill construction, accumul…
Theory-optimal Quantization Based on Flatness
Xiusheng Huang, Zhe Li, Xuanwu Yin +5
Post-training quantization has emerged as a widely adopted technique for compressing and accelerating the inference of Large Language Models (LLMs). The primary challenges in LLMs…
Mutual Enhancement Between Global Tokens and Patch Tokens: From Theory to Practice
Xiusheng Huang, Xin Jiang, Jun Zhao +2
Accurate and effective discrete image tokenization is crucial for long image sequence processing. However, current methods rigidly compress all content at a fixed rate, ignoring th…
Exploiting Contextual Knowledge in LLMs through V-usable Information based Layer Enhancement
Xiaowei Yuan, Zhao Yang, Ziyang Huang +5
Large Language Models (LLMs) have demonstrated remarkable capabilities in various tasks, yet they often struggle with context-faithfulness generations that properly reflect context…
If an LLM Were a Character, Would It Know Its Own Story? Evaluating Lifelong Learning in LLMs
Siqi Fan, Xiusheng Huang, Yiqun Yao +6
Large language models (LLMs) can carry out human-like dialogue, but unlike humans, they are stateless due to the superposition property. However, during multi-turn, multi-agent int…
Capability Localization: Capabilities Can be Localized rather than Individual Knowledge
Xiusheng Huang, Jiaxiang Liu, Yequan Wang +2
Large scale language models have achieved superior performance in tasks related to natural language processing, however, it is still unclear how model parameters affect performance…
Commonsense Knowledge Editing Based on Free-Text in LLMs
Xiusheng Huang, Yequan Wang, Jun Zhao +1
Knowledge editing technology is crucial for maintaining the accuracy and timeliness of large language models (LLMs) . However, the setting of this task overlooks a significant port…
Reasons and Solutions for the Decline in Model Performance after Editing
Xiusheng Huang, Jiaxiang Liu, Yequan Wang +1
Knowledge editing technology has received widespread attention for low-cost updates of incorrect or outdated knowledge in large-scale language models. However, recent research has…
Improving Zero-shot LLM Re-Ranker with Risk Minimization
Xiaowei Yuan, Zhao Yang, Yequan Wang +2
In the Retrieval-Augmented Generation (RAG) system, advanced Large Language Models (LLMs) have emerged as effective Query Likelihood Models (QLMs) in an unsupervised way, which re-…
Discerning and Resolving Knowledge Conflicts through Adaptive Decoding with Contextual Information-Entropy Constraint
Xiaowei Yuan, Zhao Yang, Yequan Wang +3
Large language models internalize enormous parametric knowledge during pre-training. Concurrently, realistic applications necessitate external contextual knowledge to aid models on…