Search papers, labs, and topics across Lattice.
1
0
3
LLMs can act as effective action-level supervisors in reinforcement learning, dramatically boosting the sample efficiency of SAC without sacrificing convergence guarantees.
