This dataset contains the payoff tables used in the within-subjects experiment reported in "From Overload to Convergence: Supporting Multi-Issue Human-AI Negotiation with Bayesian Visualization" (CHI 2026, Best Paper Award). It comprises two files: payoff_issues.csv: 16 negotiation issues used in a property rental scenario, with total payoff sums for both the human and AI parties. payoff_options.csv: 7 discrete options per issue (112 rows total), with individual human and AI payoffs for each option, enabling reconstruction of the full asymmetric payoff matrices. The scenario is fully integrative: human and AI preferences diverge across issues, enabling mutually beneficial trade-offs. Designed for replication of 1-, 3-, 5-, or 7 issue negotiation conditions as described in the paper.