Reproduce “Can Large Language Models Extend Their Agentic Capabilities to Goal Pursuit in Open-Ended Multi-Turn Interaction?” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Mingyu Chen, Jingkai Lin, Xiaofen Xing, Liankai Huang, Ruiqi Xiong, Xiangmin Xu
Tool-Mediated Interaction (TMI) tests whether organizing dialogue through a model’s tool loop improves goal pursuit across turns. It turns actions into tool calls and partner replies into tool results, retaining reasoning and interaction history without training. Across four simulated environments and two models, TMI leads on all four outcomes; Qwen’s buyer margin is 0.73 versus 0.51 with thinking. Evidence comes from simulated partners.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.