Search papers, labs, and topics across Lattice.
1
0
2
12
SCOPE-RL boosts LLM accuracy by over 11% while cutting reasoning tokens significantly, reshaping how we optimize reasoning paths in AI.
