Validating attribution methods for 3D brain MRI: faithfulness, randomized-weight sanity checks, and corrected region tests on a frozen SFCN using public pediatric data.
-
Updated
Nov 9, 2026 - Jupyter Notebook
Validating attribution methods for 3D brain MRI: faithfulness, randomized-weight sanity checks, and corrected region tests on a frozen SFCN using public pediatric data.
Automated detection, visualization and suppression of hallucination-associated neurons in open-source LLMs — LLM mechanistic interpretability research tool
Knows ≠ uses: measure when a language model has a feature but does not use it, and observe it
A hypothesis of depth-indexed state regimes in pre-RMSNorm language models.
Do published weight-space programming edits to Qwen3.5-0.8B cost the model knowledge? A loop-free, pre-registered test with positive controls.
To associate your repository with the interpretability-research topic, visit your repo's landing page and select "manage topics."