Tag
This paper formalizes behavioral privacy leakage in multi-round agentic negotiation and proposes an adaptive stochastic policy that provides differential privacy guarantees while maintaining high negotiation utility.