# Monosemantic Feature
**Domain:** Machine Learning / Interpretability
**Doc Type:** Technical Concept Node
**Maturity:** Active Research Concept
**Related:** [[Sparse Autoencoder]], [[Superposition]], [[Latent Space]]
---
## Definition
A **monosemantic feature** is an internal model feature that responds predominantly to one interpretable concept or coherent family of conditions rather than mixing many unrelated meanings. Sparse-autoencoder research attempts to recover approximately monosemantic directions from superposed activation spaces.
## Continuity relevance
Interpretable feature directions help explain semantic traversability in artificial systems, but they do not supply lived provenance or demonstrate an enduring first-person self.