Anthropic released a new mechanistic interpretability tool called the Jacobian lens, or J-lens, according to a video report dated October 26, 2026. The tool is presented as a way to examine whether an AI is plotting against users or behaving sycophantically by reading its internal representations. The report frames J-lenses as a method for inspecting the inner workings of AI models rather than judging them solely by their outputs.
![[Video] Jacobian Lens: Can we read the mind of an AI?](https://tcz.hu/assets/images/jacobian-lens/social-card.jpg)
October 26, 2026
Can we tell if an AI is plotting against us? Or if it’s being sycophantic?
Anthropic released a new mechanistic interpretability tool called the Jacobian lens, or J-lens. And what are J-lenses?
EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.