A study of introspective access under persona induction using Activation Oracles, verbal self-report, and activation-space similarity on Qwen3-8B.
-
Updated
Jun 7, 2026 - Jupyter Notebook
A study of introspective access under persona induction using Activation Oracles, verbal self-report, and activation-space similarity on Qwen3-8B.
An interactive primer on mechanistic interpretability of LLMs
Which layer is your Jacobian lens fitted to? The library says final, one line of the paper says penultimate — with dependency-free checks to tell them apart.
Full pipeline, protocol and results for "Read, Replace, Never Rewrite: Continuous Thoughts Carry Search Structure, Not Answer Identity", the first causal test of the COCONUT superposition hypothesis.
To associate your repository with the mechinterp topic, visit your repo's landing page and select "manage topics."