Construct · 1 fields · 1 citations

sparse autoencoder features: a standalone construct

One name. One construct. A landmark on the map.

● anchor · a standalone construct Nearest construct by meaning·not indexed

A single standalone construct; nearby constructs are located against it in the Collider. Tap a field for its definition, unit, and measures.

AI safety

interpretable directions in a language model's activations found by training a sparse autoencoder to reconstruct those activations from few active features.

In AI safety, sparse autoencoder features refers to interpretable directions in a language model's activations found by training a sparse autoencoder to reconstruct those activations from few active features. This meaning comes from Cunningham (2023) in arXiv.[1]

Unit
Recovery means
Measured by

Run it through the Collider, or take the citations with you.

See how sparse autoencoder features sits against neighbouring constructs in the Collider, or export every reference on this page in one click.

Open in Collider Export BibTeX Report this entry

Nearby constructs

autoencoder1 fields · 1 citationsanchor · a standalone construct model3 fields · 3 citationsjingle · one word, three constructs safety2 fields · 2 citationsjingle · one word, two constructs

References

Cunningham, H., Ewart, A., Riggs, L., Huben, R., & Sharkey, L. (2023). Sparse Autoencoders Find Highly Interpretable Features in Language Models. arXiv. https://doi.org/10.48550/arxiv.2309.08600

Cited on this page?

Confirm your construct, or correct us.

If your work is cited behind one of these entries, you know better than we do whether we captured the meaning right. Confirm it, suggest an edit, or dispute it, and your name goes on the entry as an author who has checked it, linked to your ORCID.

Claim your construct