Search papers, labs, and topics across Lattice.
This paper introduces BOND, a framework for training negotiation agents that explicitly model and output beliefs about their opponent's preferences. A Bayesian teacher model scores dialogue contexts against possible opponent priority orderings and updates a posterior, which is then distilled into a smaller student LM that emits both actions and normalized posterior beliefs. Experiments on the CaSiNo dataset show BOND outperforms the state-of-the-art and achieves strong Brier scores for opponent-priority posteriors, demonstrating effective distillation of Bayesian belief states.
You can distill interpretable Bayesian reasoning about opponent preferences into an 8B language model, outperforming much larger models and enabling detailed auditability of negotiation strategies.
Negotiation agents must infer what their counterpart values, update those beliefs over dialogue turns, and choose actions under uncertainty. End-to-end large language models (LLMs) can imitate negotiation dialogue, but their opponent beliefs are usually implicit and difficult to inspect. We propose BOND (Bayesian Opponent-belief Negotiation Distillation), a framework for auditable negotiation. BOND consists of an LLM-based Bayesian teacher that scores dialogue contexts against the six possible opponent priority orderings, updates a posterior over those orderings, and uses the posterior for menu-based decision making, as well as a smaller 8B student language model that emits both negotiation actions and normalized posterior beliefs as tagged text. In the CaSiNo negotiation dataset, BOND outperforms the state-of-the-art and achieves mean Brier score 0.085 over opponent-priority posteriors. The distilled student preserves much of this belief signal, achieving Brier 0.114, below the uniform six-ordering reference of 5/36, approximately 0.139. Compared with a 70B structured-CoT baseline, the significantly smaller 8B student model yields substantially better elicited posterior calibration. We further showcase auditability through posterior trajectories, belief-versus-policy error decomposition, and posterior-prefix interventions. These diagnostics reveal that distillation preserves a scoreable belief report more strongly than causal belief-conditioned control, making weak belief-action coupling visible, not hidden.
