Preserving Pre-trained Features Helps Calibrate Fine-tuned Language Models.
Guande HeJianfei ChenJun ZhuPublished in: CoRR (2023)
Keyphrases
- language model
- fine tuned
- language modeling
- n gram
- pre trained
- probabilistic model
- retrieval model
- speech recognition
- feature extraction
- fine tuning
- statistical language models
- document retrieval
- query expansion
- language modelling
- test collection
- feature vectors
- general purpose
- training data
- information retrieval
- language models for information retrieval
- statistical language modeling
- relevance model
- multi modal
- image features
- learning algorithm