Verbalized internal states detect hidden manipulations in medical LLMs
arXiv — cs.AI · Frederik Hauke, Jan Clusmann, Hanna Kreutzer, Christiane Kuhl, Tianyu Han, Felix Barajas Ordonez, Sven Nebelung, Jakob Nikolas Kather, Daniel Truhn · Published on Readzer
Summary
A cross-type arXiv study (2611.24994) reports that translating a multimodal LLM's (Gemma-3-27B) internal activations into text via a natural-language autoencoder—termed the 'thoughts channel'—caught injected attacks significantly more often than the model's answer across all eight injected conditions (p<0.05), with near-zero false positives on clean inputs. The omission result replicated on Llama-3.3-70B, and the channel flagged unsafe clinical requests before refusal decisions, requiring no retraining.
This is a reviewed summary. Read the complete article at the original source.
Why it matters by industry
- Artificial intelligence
Relevant to AI model safety and interpretability, particularly for clinical applications where hidden manipulations could bypass output-level defenses.