CLT-Forge
A Mechanistic Interpretability Toolkit for Cross-Layer Transcoder Training and Attribution-Graph Visualization
- 111
- Python
- MIT
- 4 mo ago
A Mechanistic Interpretability Toolkit for Cross-Layer Transcoder Training and Attribution-Graph Visualization
Automates attribution-graph analysis via probe prompting: circuit-trace a prompt, auto-generate concept probes, profile feature activations, cluster supernodes.
Do attribution graphs predict what GPT-2 really does? 8,064 feature edits, why graphs mislead (frozen LayerNorm/attention), a fix that keeps them linear, a trust score, a CLI and a browser demo.
Hierarchical Sparse Circuit Extraction from Billion-Parameter Language Models through Scalable Attribution Graph Decomposition
Interactive lab for attribution graphs & circuit tracing — replace polysemantic neurons with monosemantic features, trace causal paths, and patch a real two-hop reasoning circuit. MIT Tech Review 2026 Breakthrough tech.