Search papers, labs, and topics across Lattice.
Cornell University
3
0
7
MANTA reveals that multi-agent systems can dynamically adapt their communication structures during execution, leading to significant performance gains.
LLMs are often overconfident chatbots, not helpful assistants: they can spot ambiguous questions, but almost never ask for clarification before answering.
Attention from your LLM can be used to significantly improve preference optimization, outperforming existing methods without needing a separate reward model or heuristic token weighting.
