On the Algorithmic Bias of Aligning Large Language Models with RLHF: Preference Collapse and Matching Regularization.
Jiancong XiaoZiniu LiXingyu XieEmily J. GetzenCong FangQi LongWeijie J. SuPublished in: CoRR (2024)
Keyphrases
- language model
- language modeling
- probabilistic model
- n gram
- query expansion
- document retrieval
- information retrieval
- retrieval model
- speech recognition
- language modelling
- test collection
- ad hoc information retrieval
- statistical language models
- vector space model
- language models for information retrieval
- context sensitive
- translation model
- active learning
- document ranking
- smoothing methods
- pseudo relevance feedback
- query terms
- retrieval effectiveness
- okapi bm
- web documents
- spoken term detection