concept
active
concept:feature-driftFeature drift
Decoder direction of a feature changing across layers.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Accumulation of mismatch in later layers causing S degradation.
- Behavioural drift in multi-turn LLM interaction; documented in prior work for persona, identity, and instruction-following
- Domain of techniques for constructing informative features from raw data; covariance pooling is a feature engineering method for token sequences.
- Interpretable concept representations hypothesized as basic units of computation.
- Property that features activate on only a small fraction of inputs; enables compressed sensing and is what allows superposition to work
- Phenomenon where a feature in a small SAE splits into multiple finer features in a larger SAE.
- Modifying SAE hidden activations h to h-tilde while keeping error term fixed, then continuing inference.
- Repetitive behavioral pattern observed under high steering strengths in SAE feature self-steering experiments