# Persona Vectors: Monitoring and Controlling Character Traits in Language Models

arXiv: https://arxiv.org/abs/2507.21509 · HTML: https://mlsanity.com/paper/2507.21509/

## Key facts

- Cited by 345 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 29 Jul 2025.

## Built on by (most cited first, 30 of 345)

- [School of Reward Hacks: Hacking harmless tasks generalizes to misaligned behavior in LLMs](https://mlsanity.com/paper/2508.17511/) (2025)
- [Emotion Concepts and their Function in a Large Language Model](https://mlsanity.com/paper/2604.07729/) (2026)
- [Inoculation Prompting: Eliciting traits from LLMs during training can suppress them at test-time](https://mlsanity.com/paper/2510.04340/) (2025)
- [Inoculation Prompting: Instructing LLMs to misbehave at train-time improves test-time alignment](https://mlsanity.com/paper/2510.05024/) (2025)
- [Sycophancy Is Not One Thing: Causal Separation of Sycophantic Behaviors in LLMs](https://mlsanity.com/paper/2509.21305/) (2025)
- [The Personality Illusion: Revealing Dissociation Between Self-Reports & Behavior in LLMs](https://mlsanity.com/paper/2509.03730/) (2025)
- [Peacemaker or Troublemaker: How Sycophancy Shapes Multi-Agent Debate](https://mlsanity.com/paper/2509.23055/) (2025)
- [Belief Dynamics Reveal the Dual Nature of In-Context Learning and Activation Steering](https://mlsanity.com/paper/2511.00617/) (2025)
- [Narrow Finetuning Leaves Clearly Readable Traces in Activation Differences](https://mlsanity.com/paper/2510.13900/) (2025)
- [Locate, Steer, and Improve: A Practical Survey of Actionable Mechanistic Interpretability in Large Language Models](https://mlsanity.com/paper/2601.14004/) (2026)
- [Persistent Instability in LLM's Personality Measurements: Effects of Scale, Reasoning, and Conversation History](https://mlsanity.com/paper/2508.04826/) (2025)
- [Manifold Steering Reveals the Shared Geometry of Neural Network Representation and Behavior](https://mlsanity.com/paper/2605.05115/) (2026)
- [Open Character Training: Shaping the Persona of AI Assistants through Constitutional AI](https://mlsanity.com/paper/2511.01689/) (2025)
- [Sycophancy Hides Linearly in the Attention Heads](https://mlsanity.com/paper/2601.16644/) (2026)
- [How RLHF Amplifies Sycophancy](https://mlsanity.com/paper/2602.01002/) (2026)
- [Steering Language Models with Weight Arithmetic](https://mlsanity.com/paper/2511.05408/) (2025)
- [Model Spec Midtraining: Improving How Alignment Training Generalizes](https://mlsanity.com/paper/2605.02087/) (2026)
- [Subliminal Effects in Your Data: A General Mechanism via Log-Linearity](https://mlsanity.com/paper/2602.04863/) (2026)
- [Simulated Students in Tutoring Dialogues: Substance or Illusion?](https://mlsanity.com/paper/2601.04025/) (2026)
- [Emergent Misalignment via In-Context Learning: Narrow in-context examples can produce broadly misaligned LLMs](https://mlsanity.com/paper/2510.11288/) (2025)
- [Consistency Training Helps Stop Sycophancy and Jailbreaks](https://mlsanity.com/paper/2510.27062/) (2025)
- [Observing and Controlling Features in Vision-Language-Action Models](https://mlsanity.com/paper/2603.05487/) (2026)
- [What Counts as AI Sycophancy? A Taxonomy and Expert Survey of a Fragmented Construct](https://mlsanity.com/paper/2605.21778/) (2026)
- [In-Training Defenses against Emergent Misalignment in Language Models](https://mlsanity.com/paper/2508.06249/) (2025)
- [TwinVoice: A Multi-dimensional Benchmark Towards Digital Twins via LLM Persona Simulation](https://mlsanity.com/paper/2510.25536/) (2025)
- [Spilling the Beans: Teaching LLMs to Self-Report Their Hidden Objectives](https://mlsanity.com/paper/2511.06626/) (2025)
- [Recontextualization Mitigates Specification Gaming without Modifying the Specification](https://mlsanity.com/paper/2512.19027/) (2025)
- [Subliminal Steering: Stronger Encoding of Hidden Signals](https://mlsanity.com/paper/2604.25783/) (2026)
- [The Era of Real-World Human Interaction: RL from User Conversations](https://mlsanity.com/paper/2509.25137/) (2025)
- [EasySteer: A Unified Framework for High-Performance and Extensible LLM Steering](https://mlsanity.com/paper/2509.25175/) (2025)

## Abstract (arXiv)

Large language models interact with users through a simulated 'Assistant' persona. While the Assistant is typically trained to be helpful, harmless, and honest, it sometimes deviates from these ideals. In this paper, we identify directions in the model's activation space-persona vectors-underlying several traits, such as evil, sycophancy, and propensity to hallucinate. We confirm that these vectors can be used to monitor fluctuations in the Assistant's personality at deployment time. We then apply persona vectors to predict and control personality shifts that occur during training. We find that both intended and unintended personality changes after finetuning are strongly correlated with shifts along the relevant persona vectors. These shifts can be mitigated through post-hoc intervention, or avoided in the first place with a new preventative steering method. Moreover, persona vectors can be used to flag training data that will produce undesirable personality changes, both at the dataset level and the individual sample level. Our method for extracting persona vectors is automated and can be applied to any personality trait of interest, given only a natural-language description.