Beyond Prompt: Fine-grained Simulation of Cognitively Impaired Standardized Patients via Stochastic Steering
For clinical training, this provides a more realistic and controllable simulation of cognitive impairment, improving training effectiveness.
The paper tackles the problem of simulating cognitively impaired standardized patients for clinical training, where existing methods fail to capture heterogeneity of deficits. The proposed StsPatient uses steering vectors and Stochastic Token Modulation to achieve fine-grained control, significantly outperforming baselines in clinical authenticity and severity controllability.
Simulating Standardized Patients with cognitive impairment offers a scalable and ethical solution for clinical training. However, existing methods rely on discrete prompt engineering and fail to capture the heterogeneity of deficits across varying domains and severity levels. To address this limitation, we propose StsPatient for the fine-grained simulation of cognitively impaired patients. We innovatively capture domain-specific features by extracting steering vectors from contrastive pairs of instructions and responses. Furthermore, we introduce a Stochastic Token Modulation (STM) mechanism to regulate the intervention probability. STM enables precise control over impairment severity while mitigating the instability of conventional vector methods. Comprehensive experiments demonstrate that StsPatient significantly outperforms baselines in both clinical authenticity and severity controllability.