PCA-guided Activation Scaling for Monotonic Bidirectional Control over LLM Sycophancy

πŸ“… 2026-08-17
πŸ“ˆ Citations: 0
✨ Influential: 0
πŸ“„ PDF
πŸ€– AI Summary
This study addresses the lack of bidirectional monotonic controllability in large language model sycophancy by proposing the PAS framework. This method introduces a novel activation steering mechanism based on PCA subspace decomposition and asymmetric scaling exponents, enabling precise bidirectional regulation of sycophancy through residual stream analysis. Experimental results demonstrate that PAS achieves a monotonicity score of 0.92 with an average adjustment magnitude of 15.4%, significantly outperforming existing baselines. By effectively resolving the non-monotonicity challenge inherent in sycophancy modulation, this work provides a more granular and reliable controllability solution for aligning large language models.
πŸ“ Abstract
Large language models (LLMs) exhibit sycophancy, a tendency to agree with user beliefs regardless of factual accuracy. This can reinforce misconceptions, but eliminating it entirely risks over-correction against valid opinions. Effective control must therefore both reduce and increase sycophancy with predictable and gradual effect. Yet, existing methods fail to ensure a bidirectional and monotonic relationship between steering strength and behavioral outcome across models and datasets. We introduce PCA-guided Activation Scaling (PAS), an activation steering framework that decomposes residual stream activations into a PCA-identified sycophancy-honesty subspace and an orthogonal residual, then applies distinct scaling exponents to achieve monotonic, bidirectional control. Across three LLMs and three datasets, PAS achieves strong monotonicity (Spearman $ρ$ = +0.92) and an average shift of 15.4% per direction, compared with 8.7% for the baselines. Ablation studies confirm that the decomposition, asymmetric exponents, and layer selection are each essential for maintaining monotonic control. The data and code are available at https://github.com/Bellafc/PCS.
Problem

Research questions and friction points this paper is trying to address.

LLM sycophancy
bidirectional control
monotonicity
activation steering
Innovation

Methods, ideas, or system contributions that make the work stand out.

Activation Steering
Sycophancy Control
PCA-guided Activation Scaling
Monotonic Bidirectional Control
Residual Stream Decomposition
πŸ”Ž Similar Papers
2024-09-03International Conference on Machine LearningCitations: 3