Search papers, labs, and topics across Lattice.
1
0
2
19
LLM agent reliability metrics hide a wealth of information: modeling execution traces as Markov chains reveals the underlying success-time distribution and quantifies uncertainty, offering a richer understanding of agent behavior.
