PLaD: Preference-based Large Language Model Distillation with Pseudo-Preference Pairs.
Rongzhi ZhangJiaming ShenTianqi LiuHaorui WangZhen QinFeng HanJialu LiuSimon BaumgartnerMichael BenderskyChao ZhangPublished in: CoRR (2024)
Keyphrases
- language model
- language modeling
- user preferences
- preference relations
- n gram
- document retrieval
- probabilistic model
- information retrieval
- retrieval model
- query expansion
- test collection
- speech recognition
- language modelling
- pairwise
- document length
- mixture model
- smoothing methods
- ad hoc information retrieval
- statistical language models
- vector space model
- pseudo relevance feedback
- query terms
- context sensitive
- relevance model
- word clouds
- word error rate
- model selection
- search engine