**arXiv ID:** 2312.03096 **Authors:** Victor Lecomte, Kushal Thaman, Rylan Schaeffer, Naomi Bashkansky, Trevor Chow, Sanmi Koyejo **Published:** 2023-12-05T19:29:54Z **Abstract:** Polysemantic neurons -- neurons that activate for a set of unrelated features -- have been seen as a significant obstacle towards interpretability of task-optimized deep networks, with implications for AI safety. The classic origin story of polysemanticity is that the data contains more ``features" than neurons, suc...
Scanned 9/11/2026
Install to Claude Code
npx -y skills add hiyenwong/ai_collection --skill what-causes-polysemanticity-an-alternative-origin-story-of-mixed-selectivity-from-incidental-causes --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of What Causes Polysemanticity An Alternative Origin Story Of Mixed Selectivity From Incidental Causes?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/hiyenwong-what-causes-polysemanticity-an-alternative-origin)More formats (shields.io, HTML) on the badges page.
# What Causes Polysemanticity? An Alternative Origin Story of Mixed Selectivity from Incidental Causes
**arXiv ID:** 2312.03096
**Authors:** Victor Lecomte, Kushal Thaman, Rylan Schaeffer, Naomi Bashkansky, Trevor Chow, Sanmi Koyejo
**Published:** 2023-12-05T19:29:54Z
**Abstract:**
Polysemantic neurons -- neurons that activate for a set of unrelated features -- have been seen as a significant obstacle towards interpretability of task-optimized deep networks, with implications for AI safety. The classic origin story of polysemanticity is that the data contains more ``features" than neurons, such that learning to perform a task forces the network to co-allocate multiple unrelated features to the same neuron, endangering our ability to understand networks' internal processing. In this work, we present a second and non-mutually exclusive origin story of polysemanticity. We show that polysemanticity can arise incidentally, even when there are ample neurons to represent all features in the data, a phenomenon we term \textit{incidental polysemanticity}. Using a combination of theory and experiments, we show that incidental polysemanticity can arise due to multiple reasons including regularization and neural noise; this incidental polysemanticity occurs because random initialization can, by chance alone, initially assign multiple features to the same neuron, and the training dynamics then strengthen such overlap. Our paper concludes by calling for further research quantifying the performance-polysemanticity tradeoff in task-optimized deep neural networks to better understand to what extent polysemanticity is avoidable.
## Skill Description
This skill is generated from the arXiv paper: What Causes Polysemanticity? An Alternative Origin Story of Mixed Selectivity from Incidental Causes (2312.03096).
## How to Use
[To be filled in by the user or by future automation]
## References
- [arXiv:2312.03096](http://arxiv.org/abs/2312.03096v3)
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!