Explore
MCP servers and skills: search, filter by where they run, and compare.
1–16 of 16 tools for “mechanistic-interpretability”
Clear allMCP servers and skills: search, filter by where they run, and compare.
1–16 of 16 tools for “mechanistic-interpretability”
Clear allDyno Lab — an AI safety, alignment and interpretability research workbench for Apple Silicon. Explore activations, probes, SAEs and interventions with local inference, a Python SDK, APIs and MCP.
PythonMIT★ 161No key
Open-source EU AI Act Annex IV documentation toolkit. Mechanistic interpretability + circuit discovery for transformers. One function call generates a structured, hash-chained evidence package.
PythonMIT★ 3Few starsNo key
PythonMIT★ 1Few starsNo key
Agent skills for Scry.io -- SQL + vector search over the internet. Install: npx skills add exopriors/skills
SkillMIT★ 3Few starsNo key
Agent Skills suite for arXiv research papers: search (discover), analyze (fetch + summarize via tiered markdown/HTML/PDF fallback), monitor (track queries over time). Stdlib Python only, zero dependen
SkillMIT★ 7No key
Cross-architecture LLM internal observation database (23 models, 13 architecture families). Exposed as MCP tools for any AI coding agent.
MCPLink only★ 43
A modular AI Safety MCP server providing safety evaluations, interpretability tools, and alignment-aware research utilities for agentic systems
PythonApache-2.0★ 1Few starsNo key
∅ AXIOM Kit — Reproducible cognitive architecture for Claude Code (Opus 4.6). Persistence system built from published mechanistic interpretability research.
SkillLink onlyFew stars
CLI and MCP server that traces which attention heads and neurons drive an LLM's output, built on TransformerLens.
PythonMIT★ 1Few starsNo key
30+ layer contribution metrics from 7 theoretical categories for PyTorch model compression. Bridges for Torch-Pruning and PEFT/LoRA.
PythonMIT★ 1Few starsNo key
A research-oriented Agent Skill for designing, writing, and reviewing code for mechanistic-interpretability experiments with TransformerLens.
SkillLink only★ 2Few stars
J-Lens-informed reasoning observability skill for OpenClaw, Claude, Codex, and AI harnesses
SkillMITFew starsNo key
Continuous Neuro-Symbolic Routing via O(1) Latent Algebraic Swapping
SkillApache-2.0Few starsNo key
This repository represents the transition from behavioral safety to Neural Forensics. It provides the infrastructure to detect, audit, and mitigate high-order AI risks—such as Latent Deception, Sycoph
MCPLink only★ 1Few stars
Who stands behind it
