SOURCE-LINKED INTELLIGENCE

Curved Inference II: Sleeper Agent Geometry - Extending Interpretability Beyond Probes

arXiv · AI, language, vision and robotics · article · Aug 25, 2026 · UTC

This paper extends Anthropic's Sleeper Agents research [1], which showed artificial backdoors persist through safety training & can be detected by linear probes with >99% accuracy [2]. However, probe-based detection relies on linear separability that may be an artefact of backdoor insertion rather than a property of naturally occurring deceptive alignment. Sophisticated deceptive behaviours emerging through natural training are unlikely to produce such convenient linear signals. We introduce a naturalistic methodology using multi-turn context windows that simulates realistic deceptive reasonin

Read original source ↗ Open in workspace

recordType

paper

region

Global