Probing Toxic Content in Large Pre-Trained Language Models.
Nedjma OusidhoumXinran ZhaoTianqing FangYangqiu SongDit-Yan YeungPublished in: ACL/IJCNLP (1) (2021)
Keyphrases
- language model
- pre trained
- language modeling
- probabilistic model
- n gram
- retrieval model
- speech recognition
- document retrieval
- language modelling
- information retrieval
- statistical language models
- query expansion
- test collection
- document ranking
- smoothing methods
- language models for information retrieval
- multimedia
- training examples
- active learning
- training data
- semi supervised learning
- generative model
- data sets