Persona Alchemy: Designing, Evaluating, and Implementing Psychologically-Grounded LLM Agents for Diverse Stakeholder Representation
This work addresses the challenge of representing diverse stakeholders in LLM personas for applications like renewable energy discourse, offering improved explainability and reproducibility, though it is incremental in building on existing persona design efforts.
The authors tackled the problem of aligning LLM personas with human cognitive processes and diverse stakeholder perspectives by introducing a Social Cognitive Theory (SCT) agent design framework, resulting in consistent response patterns (R² range: 0.58-0.61) and systematic temporal development validated by principal component analysis explaining 73% of variance.
Despite advances in designing personas for Large Language Models (LLM), challenges remain in aligning them with human cognitive processes and representing diverse stakeholder perspectives. We introduce a Social Cognitive Theory (SCT) agent design framework for designing, evaluating, and implementing psychologically grounded LLMs with consistent behavior. Our framework operationalizes SCT through four personal factors (cognitive, motivational, biological, and affective) for designing, six quantifiable constructs for evaluating, and a graph database-backed architecture for implementing stakeholder personas. Experiments tested agents' responses to contradicting information of varying reliability. In the highly polarized renewable energy transition discourse, we design five diverse agents with distinct ideologies, roles, and stakes to examine stakeholder representation. The evaluation of these agents in contradictory scenarios occurs through comprehensive processes that implement the SCT. Results show consistent response patterns ($R^2$ range: $0.58-0.61$) and systematic temporal development of SCT construct effects. Principal component analysis identifies two dimensions explaining $73$% of variance, validating the theoretical structure. Our framework offers improved explainability and reproducibility compared to black-box approaches. This work contributes to ongoing efforts to improve diverse stakeholder representation while maintaining psychological consistency in LLM personas.