Search papers, labs, and topics across Lattice.
1
0
3
You can now audit multi-agent LLM systems and trace responsibility for harmful outputs even without access to internal execution logs, thanks to a clever "self-describing text" technique.
