ConceptAI & ML
Logits
The raw, unnormalized output values of a model before applying a softmax function.
Know first
Named as a prerequisite by the lectures that define this. The number is how many say so. A grey one has no page yet.
Opens up
Concepts that name this one as a prerequisite.
Lecturers
2Each in their own terms, from the easiest level.
Andrej Karpathy
Raw, unnormalized outputs of a neural network layer interpreted as log counts.
Lectures that assume it
5Connections
14The lectures that define it or assume it, and the concepts next to it.
Concepts7
Show 1 moreShow fewer
Lectures7
- Spelled-Out Intro to Language Modeling: Building makemore (Bigram & Cross-Entropy)
- Neural Networks I: Feedforward Architectures & Backpropagation
- Building makemore Part 3: Activations, Gradients, Kaiming Init & BatchNorm
- Let's Reproduce GPT-2 (124M): Distributed PyTorch, FlashAttention & FineWeb
- Just Machine Learning
- GSQ-RCO: Better Non-Uniform GGUF Quantization