Analyzing Feed-Forward Blocks in Transformers through the Lens of Attention Maps.
Goro KobayashiTatsuki KuribayashiSho YokoiKentaro InuiPublished in: ICLR (2024)
Keyphrases
- feed forward
- back propagation
- neural network
- artificial neural networks
- neural nets
- biologically plausible
- recurrent neural networks
- feed forward neural networks
- visual cortex
- activation function
- neural architecture
- artificial neural
- hidden layer
- error back propagation
- primate visual cortex
- fractal image coding
- spiking neural networks
- data mining
- recurrent networks
- neuron model