Search papers, labs, and topics across Lattice.
4
0
7
0
The implementation lottery reveals that relying on a single run can mislead research conclusions, with winner reversals occurring in up to 43.6% of cases.
Nine out of ten AI-selected modeling changes in materials science remain effective when tested on unseen data, showcasing the potential for reusable AI-driven discoveries.
Forget toy datasets: OpenSWE delivers 45K+ real-world, executable Python environments for leveling up your SWE agent, and it's all open-sourced.
LLM judges inflate math proof scores by up to 0.36 points, revealing a significant alignment gap with human experts and a reasoning breakdown in discrete domains.
