- Rudolf Peierls Centre for Theoretical Physics, University of Oxford
- Leinweber Institute for Theoretical Physics, Stanford University
APP-260814-0000mechanistic-interpretabilityRelease v1.0.0Listed Aug 14, 2026
Open in your agent
git clone --recurse-submodules --branch v1.0.0 https://github.com/lccqqqqq/sae-feature-nonlocality.gitcd sae-feature-nonlocalityclaudegit clone --recurse-submodules --branch v1.0.0 https://github.com/lccqqqqq/sae-feature-nonlocality.gitcd sae-feature-nonlocalitycodexThese commands clone the repository and initialize the AI agent in its directory. Ask the agent to explain methodology, clarify derivations, or reproduce figures. The repository supports Claude Code, Codex, and any agent that reads AGENTS.md.
To import this paper into another project, install the APP plugin and run /load-paper https://github.com/lccqqqqq/sae-feature-nonlocality/releases/tag/v1.0.0.
Summary
Sparse autoencoders (SAEs) decompose language-model activations into features, but knowing what a feature responds to (its description) does not settle at what level of abstraction it operates — a token-matching feature and a genuinely contextual one can carry similar descriptions. The paper introduces Feature Nonlocality (FNL): for one firing of a feature, attribute the activation to the context positions that influence it, normalize those per-position influences into a distribution, and take its entropy. A feature driven by a single token has near-zero FNL; a feature integrating a whole passage has high FNL. The paper shows FNL behaves as an abstractness measure should — it rises with network depth, is stable across text corpora once a reliability ceiling is accounted for, separates contextual from token-driven features, and predicts robustness of activation under meaning-preserving paraphrase — and then uses it in two applications: auditing a published SAE-based jailbreak mitigation (finding its features are mostly positional artifacts, not detectors of harmful intent) and selecting features for steering (steering high-FNL features improves benchmark accuracy where steering low-FNL ones does not, though gains are model-specific).