Reproduce “Prospective Memory for Proactive LLM Agents” 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
Yuanchen Bei, Zhining Liu, Tianxin Wei, Weizhi Zhang, Zhichen Zeng, Philip S. Yu, Hendrik Hamann, Jingrui He and 1 more
Pro-Memory represents unfinished user intentions as pending objectives, links them to supporting evidence, and checks new observations to decide whether to act, wait, or seek information. It coordinates proposed contributions with the agent’s actions and updates intention status when actions are issued or evidence resolves them. On ProactiveBench, Claude Haiku 4.5 achieves 0.6634 F1 versus RWM’s 0.3817; these benchmark results do not establish real-world effectiveness.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.