Proxy Confidence: Auditing Black-Box LLM Agents with a Surrogate's Log-Probabilities
arXiv:2610.03894v1 Announce Type: new Abstract: A deployed LLM agent emits tool calls, queries, and code that can be silently wrong -- by the time the error surfaces, the action has run. Frontier chat APIs hide the model's token...
Read more: https://arxiv.org/abs/2610.03894
Scan with your phone
Scan with your phone