Detecting Strategic Deception Using Linear Probes — LessWrong

lesswrong.com · 1,437 words · saved by 1 readers

Can you tell when an LLM is lying from the activations? Are simple methods good enough? We recently published a paper investigating if linear probes detect when Llama is deceptive. Abstract: AI models might use deceptive strategies as part of scheming or misaligned behaviour. Monitoring outputs alone is insufficient, since the AI might produce seemingly benign outputs while its internal reasoning is misaligned. We thus evaluate if linear probes can robustly detect deception by monitoring model activations. We test two probe-training datasets, one with contrasting instructions to be honest or deceptive (following Zou et al., 2023) and one of responses to simple roleplaying scenarios. We test whether these probes generalize to realistic settings where Llama-3.3-70B-Instruct behaves deceptively, such as concealing insider trading (Scheurer et al., 2023) and purposely underperforming on safety evaluations (Benton et al., 2024). We find that our probe distinguishes honest and deceptive resp

x Detecting Strategic Deception Using Linear Probes — LessWrong Language Models (LLMs) Apollo Research (org) Interpretability (ML & AI) AI Frontpage 2025 Top Fifty: 14 % 104 Detecting Strategic Deception Using Linear Probes by Nicholas Goldowsky-Dill , bilalchughtai , StefanHex , Marius Hobbhahn 6th Feb 2025 AI Alignment Forum 2 min read 9 104 Ω 46 This is a linkpost for https://arxiv.org/abs/2502.03407 Can you tell when an LLM is lying from the activations? Are simple methods good enough? We recently published a paper investigating if linear probes detect when Llama is deceptive. Abstract: AI

Detecting Strategic Deception Using Linear Probes — LessWrong

related reading