Mitigating Hallucination in Fictional Character Role-Play
This addresses the problem of out-of-character behavior and hallucinations in role-playing applications like customer support and embodied agents, representing an incremental improvement.
The paper tackles hallucination in fictional character role-play by introducing RoleFact, a method that improves factual precision by 18% for adversarial questions and reduces temporal hallucination by 44%.
Role-playing has wide-ranging applications in customer support, embodied agents, and computational social science. The influence of parametric world knowledge of large language models (LLMs) often causes role-playing characters to act out of character and to hallucinate about things outside the scope of their knowledge. In this work, we focus on the evaluation and mitigation of hallucination in fictional character role-play. We introduce a dataset with over 2,000 characters and 72,000 interviews, including 18,000 adversarial questions. We propose RoleFact, a role-playing method that mitigates hallucination by modulating the influence of parametric knowledge using a pre-calibrated confidence threshold. Experiments show that the proposed method improves the factual precision of generated responses by 18% for adversarial questions with a 44% reduction in temporal hallucination for time-sensitive interviews. The code and the dataset are available at https://github.com/NafisSadeq/rolefact.git.