Mechanistic interpretability on Apple Silicon: steering vectors, residual capture, and SAE analysis for MLX models
Behavioral auditing toolkit for LLMs — audit any model across 8 dimensions (factual, toxicity, bias, sycophancy, reasoning, refusal, deception, over-refusal) using teacher-forced confidence probes.
Compress LLMs while auditing whether they still know truth vs myths. SVD compression + false-belief detection in one toolkit.
