Search papers, labs, and topics across Lattice.
Beijing Institute of Technology, Xiaohongshu Inc
1
0
3
LLMs can reason more efficiently by sharing intermediate thoughts during parallel search, achieving better accuracy with less computation.
