The method called J‑tér makes it possible to read, audit, and shape the current “thoughts” of the Claude language model, which helps maintain models' reliability alongside their increasing capabilities. According to the proposal, this technique shows unexpected parallels with the functioning of the human mind and can be used as an auditing and oversight tool.
The J‑tér: insight into and influence over Claude's current thoughts
The method called J‑tér makes it possible to read, audit, and shape the current “thoughts” of the Claude language model, which helps maintain models' reliability alongside their increasing capabilities.
