Search papers, labs, and topics across Lattice.
Univ. of Southern California
2
0
5
1
Learning robotic reward functions from a million trajectories reveals that comparing entire trajectories, not just individual frames, unlocks better generalization and learning from suboptimal data.
LLM editing just got a whole lot less destructive: CrispEdit slashes capability degradation to under 1% while still nailing targeted edits, blowing past previous methods.
