ToMELP: A Theory-of-Mind Benchmark for Route-Controlled Persuasion under the Elaboration Likelihood Model

Theory of Mind (ToM) is widely regarded as central to effective persuasion, yet existing evaluations often fail to capture the infer–apply loop that arises in real-world dialogue. We introduce T HEORY - OF -M IND - G UIDED E LABORATION -L IKELIHOOD P ER - SUASION , a benchmark that jointly conditions on the audience persona and the Elaboration Likelihood Model (ELM) route ( central vs. peripheral ) within persuasive conversations. The benchmark tests whether large language models can perform ToM inference over multi-turn interactions and leverage these inferences for controllable persuasive generation. T O MELP provides a structured interface with evidence annotations, enabling automated evaluation of persuasive effectiveness, route alignment/deviation, evidence quality under the central route, and robustness to perturbations.The source code is available 1 .