---
title: "Researchers Find Steering AI Personality Neurons Barely Changes What Models Say"
subtitle: "New arXiv paper probes where Big Five personality traits live inside LLMs and shows that steering internal representations does not fully steer what the model actually says."
url: "https://alphasignal.ai/news/researchers-find-steering-ai-personality-neurons-barely-changes-what-models-say"
author: "AlphaSignal Newsroom"
author_url: "https://alphasignal.ai/editorial-team"
company: "AlphaSignal"
company_url: "https://alphasignal.ai/company/alphasignal"
published: "2026-04-13T00:00:00.000Z"
updated: "2026-08-24T03:07:19.145Z"
type: "Paper"
topics:
  - "Llms"
  - "Security"
subtopics:
  - "Hallucinations"
  - "Red Teaming"
upvotes: 1105
source_url: "https://arxiv.org/abs/2604.11802"
comments: 0
truncated: true
---

# Researchers Find Steering AI Personality Neurons Barely Changes What Models Say

> New arXiv paper probes where Big Five personality traits live inside LLMs and shows that steering internal representations does not fully steer what the model actually says.

A new arXiv paper by Yuto Harada and Hiro Taiyo Hamada sits between psychology and mechanistic interpretability: if a language model can imitate a personality, where inside the network does that personality live, and can you push it around by editing a handful of neurons? Concept-selective neurons can be identified and manipulated, but the shift they produce on internal representations does not fully carry over to what the model generates, exposing a real gap between reading a concept off a model and steering its behavior.

The work, [Psychological Concept Neurons](https://arxiv.org/abs/2604.11802), targets the Big Five personality framework, the standard psychology taxonomy of Extraversion, Agreeableness, Conscientiousness, Negative Emotionality, and Open-Mindedness. Prior research has shown LLMs can role-play personalities or infer them from user text, but almost all of that evidence is behavioral. This paper goes inside the weights.

## Three questions, one questionnaire

The authors organize the study around three research questions: where trait concepts emerge across depth, which neurons are selective for each trait, and whether intervening on those neurons can causally steer outputs. As their probe data they use the Big Five Inventory 2 (BFI-2), a validated psychology questionnaire, treating domain prediction from each item as a 5-way classification task.

![Overview of the Big Five framework, BFI-2 questionnaire, and the three research questions on neural representation and causal manipulation](https://arxiv.org/html/2604.11802v1/figure/overview.png) 

Experiments run on four instruction-tuned 7 to 8B models: Meta-Llama-3-8B-Instruct, Gemma-7B-it, Mistral-7B-Instruct, and Qwen2.5-7B-Instruct, with Llama-3 used as the primary target for the fine-grained neuron work. The methodology has three moving parts:

_Preview only. Full text requires an AlphaSignal Pro account._

## Sources

- [Psychological Concept Neurons: Can Neural Control Bias Probing and Shift Generation in LLMs?](https://arxiv.org/abs/2604.11802)
- [Psychological Concept Neurons (HTML version)](https://arxiv.org/html/2604.11802v1)
- [Understanding intermediate layers using linear classifier probes](https://openreview.net/forum?id=ryF7rTqql)
- [Probing then editing response personality of large language models](https://openreview.net/forum?id=z9SbcYYPOM)
- [Finding experts in transformer models](https://arxiv.org/abs/2005.07647)