Pre-trained Large Language Models Use Fourier Features to Compute Addition.
Tianyi ZhouDeqing FuVatsal SharanRobin JiaPublished in: CoRR (2024)
Keyphrases
- language model
- language modeling
- pre trained
- language modelling
- speech recognition
- n gram
- document retrieval
- probabilistic model
- statistical language models
- information retrieval
- feature vectors
- feature extraction
- prior knowledge
- query expansion
- support vector
- retrieval model
- smoothing methods
- feature selection
- computer vision
- neural network
- feature space
- test collection
- document ranking