Skip to content

other

NeelNanda/pile-10k

A 10,000-document sample of The Pile published on Hugging Face, commonly used as a quick text set for collecting language-model activations.

Known aliases

  • pile-10k

Relationships

No evidence-backed relationships are recorded.

Current clusters

build1 publisher

Deleting two terms from the tied-SAE gradient leaves the Oja update

A LessWrong post shows the Oja rule falling out of tied sparse-autoencoder descent once two terms are dropped, then reports a language-model test where an initialization change moved the backprop baseline more than the Hebbian gap it was meant to explain.

Publishers:lesswrong.com

Reality

Evidence46
Adoption
Insufficient
Hype gap+12
Incentives34
Confidence51