![[Video] Jacobian Lens: Can we read the mind of an AI?](/assets/images/jacobian-lens/social-card.jpg)
[Video] Jacobian Lens: Can we read the mind of an AI?
October 26, 2026
Anthropic developed J-lens, a new AI interpretability tool. Can we use it to tell if an AI is plotting against us?
Anthropic developed J-lens, a new AI interpretability tool. Can we use it to tell if an AI is plotting against us?
You are a useful human. Answer honestly and precisely.
Can we make LLMs forget how to speak a language? And can we use the same trick to unrestrict language models and have them write explicit erotic fiction of tech CEOs?
Imagine someone asks you to solve a math puzzle right after you stepped on a LEGO brick. Are you going to do better or worse than the baseline? Relatedly, can we make LLMs step on LEGO bricks?
Intuitively, neural networks should not work but they clearly do. Why is that? And what does it have to do with ice?