4 papers
Direction-Flipped Influence Audits Reveal Hidden Structure in Moral Choices of LLMs
Phil Blandfort, Tushar Karayil, Alex McKenzie +3
Moral benchmarks for LLMs typically score models on context-free prompts, implicitly treating the measured choice rate as stable. We test this assumption with a direction-flipped i…
Learning Self-Interpretation from Interpretability Artifacts: Training Lightweight Adapters on Vector-Label Pairs
Keenan Pepper, Alex McKenzie, Florin Pop +6
Self-interpretation methods prompt language models to describe their own internal states, but remain unreliable due to hyperparameter sensitivity. We show that training lightweight…
Endogenous Resistance to Activation Steering in Language Models
Alex McKenzie, Keenan Pepper, Stijn Servaes +6
Large language models can recover mid-generation from task-misaligned activation steering, producing explicit verbal restarts (e.g., ``wait, that's not right'') and continuing on-t…
Detecting High-Stakes Interactions with Activation Probes
Alex McKenzie, Urja Pawar, Phil Blandfort +4
Monitoring is an important aspect of safely deploying Large Language Models (LLMs). This paper examines activation probes for detecting ``high-stakes'' interactions -- where the te…