Search papers, labs, and topics across Lattice.
1
0
3
LRMs can be made more efficient and accurate by strategically adjusting their output length based on task difficulty, leading to a better accuracy-length trade-off.
