How does Architecture Influence the Base Capabilities of Pre-trained Language Models? A Case Study Based on FFN-Wider Transformer Models.
Xin LuYanyan ZhaoBing QinPublished in: CoRR (2024)
Keyphrases
- language model
- probabilistic model
- statistical language models
- n gram
- language modeling
- relevance model
- language modelling
- information retrieval
- pre trained
- test collection
- document retrieval
- retrieval model
- translation model
- statistical language modeling
- smoothing methods
- bayesian networks
- speech recognition
- prior knowledge
- training data