General learned delegation by clones

Darren Li, Meiqi Chen, Chenze Shao +2

Frontier language models improve with additional test-time computation, but serial reasoning or uncoordinated parallel sampling can be compute-inefficient under fixed inference bud…