Empirical evaluation on the PLUME benchmark shows steering vectors vary widely in trait expressibility, degrade on task transfer, and lose effectiveness when multiple vectors are composed.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CL 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
Stable personality vectors in LLMs function as intrinsic guardrails, with ablation increasing emergent misalignment above 40% and amplification reducing it below 3%, enabling zero-shot transfer from aligned to corrupted models.
citing papers explorer
-
On the Limits of Steering Vectors for Preference-Aligned Generation
Empirical evaluation on the PLUME benchmark shows steering vectors vary widely in trait expressibility, degrade on task transfer, and lose effectiveness when multiple vectors are composed.
-
Intrinsic Guardrails: How Semantic Geometry of Personality Interacts with Emergent Misalignment in LLMs
Stable personality vectors in LLMs function as intrinsic guardrails, with ablation increasing emergent misalignment above 40% and amplification reducing it below 3%, enabling zero-shot transfer from aligned to corrupted models.