Search papers, labs, and topics across Lattice.
1
0
3
Forget reward engineering: this work shows LLMs can self-evolve and outperform larger models by learning to explore and summarize new environments autonomously.
