Search papers, labs, and topics across Lattice.
1
0
3
3
LLMs can now plan complex, long-horizon household tasks from ambiguous instructions with significantly improved performance, thanks to a novel reinforcement learning approach that corrects intermediate reasoning steps.
