conftrace_

neural network activation

5 papers

Explore in graph

Co-occurring keywords

neural network (6616) feature learning (1311) slow feature learning (1) orientation-selective features (1) activation function (133) mechanistic interpretability (133) sparse autoencoder (112) channel denoising (2) forward-backward propagation (1) activation mechanism (1)

Papers

Repetition Neurons: How Do Language Models Produce Repetitions? NAACL 2025
Dual-Perspective Activation: Efficient Channel Denoising via Joint Forward-Backward Criterion for Artificial Neural Networks NIPS 2024
Does Large Language Model Contain Task-Specific Neurons? EMNLP 2024
Identifying Functionally Important Features with End-to-End Sparse Dictionary Learning NIPS 2024
Slow, Decorrelated Features for Pretraining Complex Cell-like Networks NIPS 2009