5 papers
Stronger-MAS: Multi-Agent Reinforcement Learning for Collaborative LLMs
Yujie Zhao, Lanxiang Hu, Yang Wang +4
Multi-agent systems (MAS) and reinforcement learning (RL) are widely used to enhance the agentic capabilities of large language models (LLMs). MAS improves task performance through…
Towards Interpretable and Inference-Optimal COT Reasoning with Sparse Autoencoder-Guided Generation
Daniel Zhao, Abhilash Shankarampeta, Lanxiang Hu +2
We propose a novel method that leverages sparse autoencoders (SAEs) and clustering techniques to analyze the internal token representations of large language models (LLMs) and guid…
General Modular Harness for LLM Agents in Multi-Turn Gaming Environments
Yuxuan Zhang, Haoyang Yu, Lanxiang Hu +2
We introduce a modular harness design for LLM agents that composes of perception, memory, and reasoning components, enabling a single LLM or VLM backbone to tackle a wide spectrum…
ReFoRCE: A Text-to-SQL Agent with Self-Refinement, Consensus Enforcement, and Column Exploration
Minghang Deng, Ashwin Ramachandran, Canwen Xu +4
We present ReFoRCE, a Text-to-SQL agent that tops the Spider 2.0 leaderboard--a challenging benchmark reflecting complex, real-world Text-to-SQL scenarios. While Text-to-SQL system…
TrimLLM: Progressive Layer Dropping for Domain-Specific LLMs
Lanxiang Hu, Tajana Rosing, Hao Zhang
Specializing large language models (LLMs) for local deployment in domain-specific use cases is necessary for strong performance while meeting latency and privacy constraints. Howev…