Search papers, labs, and topics across Lattice.
Affiliation:
5
0
8
13
Wrapping standard CLI coding agents in a state-persisting orchestration layer beats unconstrained autonomous agents at research completeness without altering the underlying model.
OpenSkill enables LLM agents to autonomously evolve their skills and verification mechanisms in open-world settings, achieving superior performance without any target-task supervision.
The hardest AI tasks remain largely unsolved, with current models achieving only a 2.6% success rate on economically valuable workflows.
Despite their visual reasoning prowess, today's MLLMs still struggle to understand handwritten math scratchwork, falling far short of human expert performance in diagnosing student errors.
LLMs in medicine may be dangerously overhyped: even the best models achieve only 39% accuracy on a contamination-free, real-world clinical benchmark, with performance tanking on newer cases.
