Search papers, labs, and topics across Lattice.
1
0
3
Scaling synthetic environments with automatically generated tasks and verifiers unlocks significant reasoning improvements in language models, achieving a 27% relative gain on BBEH.
