Search papers, labs, and topics across Lattice.
3
0
5
None of the 30 LLM agents evaluated in CausalGame demonstrated reliable causal thinking, revealing a critical gap in AI's ability to perform scientific reasoning.
Traditional metrics fail to capture the true memory capabilities of LLMs, exposing a critical gap in how we assess their deployment readiness.
Training on compound reasoning traces yields better generalization than isolated atomic modules, reshaping our understanding of how LLMs can learn to reason.
