trends.barnas.net

architecture · first seen 17 Oct 2026

JumpReLU

A sparse autoencoder activation function with learned per-feature thresholds that enables sparse, differentiable features.

Mentions per week

0119 May11 Aug10 NovPapersCodeDiscussion

PapersCodeDiscussion

Earliest sources

Latest sources

Mentioned together with

sparse autoencoderTopK sparse autoencodercross-layer transcodercircuit tracing