10 papers
Test-Time Unlearning via Sparse Autoencoder
Pingzhi Li, Jinhao Duan, Vaishnav Tadiparthi +6
Machine unlearning aims to remove specific knowledge from a trained large language model (LLM) without retraining from scratch. Existing methods modify model weights via gradient a…
Generative Skill Composition for LLM Agents
Xinyu Zhao, Zhen Tan, Vaishnav Tadiparthi +5
Recent LLM agents benefit from skills for solving complex tasks. Skills encapsulate modular packages of procedural knowledge and instructions for performing specialized tasks, such…
Large language models reorganize representational geometry during in-context learning
Hua-Dong Xiong, Li Ji-An, Robert C. Wilson +2
Large language models (LLMs) show remarkable flexibility in adapting to novel tasks without parameter updates, a capacity known as in-context learning (ICL). Prior work has sought…
Understanding the Role of Hallucination in Reinforcement Post-Training of Multimodal Reasoning Models
Gengwei Zhang, Jie Peng, Zhen Tan +6
The recent success of reinforcement learning (RL) in large reasoning models has inspired the growing adoption of RL for post-training Multimodal Large Language Models (MLLMs) to en…
In-context superposition: human-like working memory interference in large language models
Hua-Dong Xiong, Li Ji-An, Jiaqi Huang +3
Intelligent systems must maintain and manipulate task-relevant information online to adapt to dynamic environments. This capacity, known as working memory, is fundamental to human…
Learning Robust Reasoning through Guided Adversarial Self-Play
Shuozhe Li, Vaishnav Tadiparthi, Kwonjoon Lee +6
Reinforcement learning from verifiable rewards (RLVR) produces strong reasoning models, yet they can fail catastrophically when the conditioning context is fallible (e.g., corrupte…
Metacognitive Self-Correction for Multi-Agent System via Prototype-Guided Next-Execution Reconstruction
Xu Shen, Qi Zhang, Song Wang +8
Large Language Model based multi-agent systems (MAS) excel at collaborative problem solving but remain brittle to cascading errors: a single faulty step can propagate across agents…
Learning Physical Interaction Skills from Human Demonstrations
Tianyu Li, Hengbo Ma, Sehoon Ha +1
Learning physical interaction skills, such as dancing, handshaking, or sparring, remains a fundamental challenge for agents operating in human environments, particularly when the a…
Symbolic Graph Inference for Compound Scene Understanding
FNU Aryan, Simon Stepputtis, Sarthak Bhagat +4
Scene understanding is a fundamental capability needed in many domains, ranging from question-answering to robotics. Unlike recent end-to-end approaches that must explicitly learn…
Language Grounded Multi-agent Reinforcement Learning with Human-interpretable Communication
Huao Li, Hossein Nourkhiz Mahjoub, Behdad Chalaki +5
Multi-Agent Reinforcement Learning (MARL) methods have shown promise in enabling agents to learn a shared communication protocol from scratch and accomplish challenging team tasks.…