Search papers, labs, and topics across Lattice.
2
0
5
17
General LLMs struggle with enzyme classification, but leveraging external knowledge can dramatically improve their performance, revealing hidden gaps in reasoning capabilities.
Shared descriptive patterns in skill libraries can obscure task-relevant signals, but SkillSight reveals and calibrates these biases to boost retrieval accuracy by over 20%.
