Story
arxiv_cs_cl · Jul 1, 2026 · paper
Source brief
QuasiMoTTo: Quasi-Monte Carlo Test-Time Scaling
arxiv.orgJul 1, 2026
original source linked
In brief
Scaling inference compute, by generating many parallel attempts per problem, is a costly but reliable lever for improving language model capabilities. By default these attempts are generated independently, wasting inf...
Feed lens
eval