In early October 2026 I shared a thread on X summarizing Emmanuel Ameisen’s recent interview about interpretability research. I highlighted the field’s nascent status, the J‑space probing experiments, and the decision to treat LLMs with anthropomorphic reasoning when forecasting behavior.
Interpretability as a field is small, young, and open to newcomers.
In one experiment, they swapped “Spanish” for “French” inside the J space.
Asked what language a passage was in, the model said French.
Asked to simply continue the passage, it kept writing Spanish.