mechinterp-state

What problem does it solve?

This skill organizes hypotheses, evidence, and activity history for mechanistic interpretability research on SAE features, enabling structured knowledge capture and progress tracking.

Core Features & Use Cases

• Create and manage hypotheses about feature behavior with confidence tracking.
• Link experimental evidence to hypotheses and maintain a growing research history.
• Generate concise summaries and export notes for documentation and sharing.
• Use in ongoing investigations to plan next steps and consolidate findings.

Quick Start

Example usage: from splatnlp.mechinterp.state import ResearchStateManager; manager = ResearchStateManager(feature_id=18712, model_type='ultra'); print(manager.get_summary())