igorbarshteyn/jlens-gguf

Companion code for the global workspace interpretability paper

J-space / Jacobian-lens visualizer for Qwen3.6-27B (4-bit) on Apple Silicon, ported to Apple MLX

Extract residual-stream activations and apply steering vectors (including activation oracles) to any vLLM model during inference.

OpenAI-compatible server with a live view into any HF model's residual stream. pip install brainscope

A library for mechanistic interpretability of GPT-style language models

Unified access to Large Language Model modules using NNsight

Steering-first neuron explorer for Qwen3.5-4B: push every MLP neuron, keep every rollout, browse the results

ICA Lens: compact ICA-based interpretability tools for exploring LLM activations. Code release for the paper.

To know what models don't say out loud.