Tag
This paper presents PTEI, a framework that integrates personality traits (MBTI and OCEAN) into LLMs to enhance emotional intelligence, using contrastive learning and personality-aware prompts. Experiments show significant improvements in emotional understanding, especially when combined with Chain-of-Thought reasoning.
This paper introduces a method for analyzing and controlling language model personalities using the OCEAN framework, training low-rank adapters to manipulate traits, and demonstrates additive composition and safety implications.
This paper examines when and why self-reported psychometric measures predict the actual behavior of large language models, finding that fine-grained, behavior-specific instruments (Theory of Planned Behavior) achieve human-level coherence within a shared conversation, while broad traits like Big 5 do not.
This paper proposes a framework using Supervised Semantic Differential to represent psychological constructs as directions in a shared word-embedding space, enabling comparison across different measurement instruments and research traditions.
This research paper investigates how human personality traits and AI design characteristics jointly impact human-AI interactions in imperfectly cooperative scenarios using both simulated datasets (2,000 simulations) and human subjects experiments (290 participants). The study finds significant divergences between simulation and real-world interactions, with AI transparency emerging as a critical factor in actual human-AI encounters.