# Sycophancy as compositions of Atomic Psychometric Traits

arXiv: https://arxiv.org/abs/2508.19316 · HTML: https://mlsanity.com/paper/2508.19316/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 5 of its 43 references that are in the map.
- First posted on arXiv 26 Aug 2025.

## Builds on (most cited first, 5 of 5)

- [Steering Llama 2 via Contrastive Activation Addition](https://mlsanity.com/paper/2312.06681/) (2023)
- [Towards Understanding Sycophancy in Language Models](https://mlsanity.com/paper/2310.13548/) (2023)
- [Refusal in Language Models Is Mediated by a Single Direction](https://mlsanity.com/paper/2406.11717/) (2024)
- [Persona Vectors: Monitoring and Controlling Character Traits in Language Models](https://mlsanity.com/paper/2507.21509/) (2025)
- [Controllable Context Sensitivity and the Knob Behind It](https://mlsanity.com/paper/2411.07404/) (2024)

## Built on by (most cited first, 1 of 1)

- [Resist and Update: Counterfactual Report Coordinates for Incentive-Compatible LLMs](https://mlsanity.com/paper/2607.12985/) (2026)

## Abstract (arXiv)

Sycophancy is a key behavioral risk in LLMs, yet is often treated as an isolated failure mode that occurs via a single causal mechanism. We instead propose modeling it as geometric and causal compositions of psychometric traits such as emotionality, openness, and agreeableness - similar to factor decomposition in psychometrics. Using Contrastive Activation Addition (CAA), we map activation directions to these factors and study how different combinations may give rise to sycophancy (e.g., high extraversion combined with low conscientiousness). This perspective allows for interpretable and compositional vector-based interventions like addition, subtraction and projection; that may be used to mitigate safety-critical behaviors in LLMs.