Reproduce “FindLink: Learning to Assess Scientific Connections from Source-Grounded Premises” 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
Jingxuan Wei, Song Tan, Xiaohan Liu, Chao Sang, Changwei Wang, Linzhuang Sun, Bihui Yu, Fuqiang Wang and 1 more
FindLink assesses what supplied findings support about a proposed connection by grounding claims in their sources and study contexts, then organizing them into a support plan that preserves unresolved conditions. Its compiler checks plan structure and source bindings, not scientific entailment. On a 1,545-question fixed-evidence test, it scores 87.25 BFM versus 67.95 for same-backbone Default SFT, within English gastrointestinal oncology.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.