Shreyagoudarigela·Oct 91. Deep Learning’s New Default: Why Transformers and LLMs Ditched ReLU for GELU
Saipallaviparupathi·Oct 7Activation Functions in Deep LearningActivation Function is one of the most important components of an artificial neural network(ANN).
Itsharshit·Sep 12The 3×3 Window That Learned to SeeHow a convolutional neural network turns raw pixels into a confident “that’s a cat” — one small sliding filter at a time.
Yasiru Raveen·Sep 6Building a Neural Network Using Numpy: Part 2Here is the link for part 1 if you didn’t read it.
InThe Thoughtful EngineerbyKobi Toueg·Jul 23AI/ML Under the Hood — Part 22: Why One Neural Network Can Have Millions of TwinsThe Hidden Mathematical Symmetries Behind Deep Learning
Vijwal Sheshadri·Jun 4Why ReLU Changed Deep Learning Forever: The Mathematics Behind Modern Neural NetworksIntroduction
nomadic_seeker·Jan 7Visualizing ReLU Networks with Topology (Thinking Out of BlackBox Why and How ReLU works)we all know we need to add non linear activation functions to neural nets to make it work otherwise it is just a Linear Regression with…
Dharmendra Pratap Singh·Dec 31, 2025What Is a Neuron? Understanding Neural Networks from ScratchNeural networks often feel intimidating because people jump straight into code and equations.
Akshat Gupta·Dec 27, 2025Why Non Linearity Matters in Neural NetworksHow layers build from edges to shapes to parts to objects
