finding
active
finding:autointerp-labels-distinguish-features-from-their-ten-nearest-neighbours-in-embedding-spaceAutointerp labels distinguish features from their ten nearest neighbours in embedding space
Shown in graph for the nearest-neighbour test.
Source paper
extracted_fromNeighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Justification of scoring.
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.
- Sparse Autoencoders Find Highly Interpretable Features in Language Models (Cunningham et al., 2023)concept0.765Core methodology paper for SAE-based interpretable feature extraction
- Central claim of the paper: the method scales to state-of-the-art transformers.
- §7.2 assertion.
- Ablation flattens experiential language.
- Activation-triggered examples plus direct logit attribution.
- Central claim of the paper, supported by detailed feature analysis, human evaluation, automated interpretability of activations, and automated interpretability of logit weights
- List experiment interpretation.
- The J-space can hold more concepts when they have coherent relationships than unrelated ones.hypothesis0.722Tested with lists.