DeCoT: Debiasing Chain-of-Thought for Knowledge-Intensive Tasks in Large Language Models via Causal Intervention.
Junda WuTong YuXiang ChenHaoliang WangRyan A. RossiSungchul KimAnup B. RaoJulian J. McAuleyPublished in: ACL (1) (2024)
Keyphrases
- language model
- knowledge intensive
- knowledge acquisition
- language modeling
- document retrieval
- probabilistic model
- information retrieval
- n gram
- speech recognition
- language modelling
- retrieval model
- law enforcement
- query expansion
- context sensitive
- smoothing methods
- statistical language models
- query terms
- document ranking
- test collection
- language models for information retrieval
- vector space model
- human resources
- relevance model
- translation model
- language model for information retrieval
- ad hoc information retrieval
- spoken term detection
- pseudo relevance feedback
- information sharing
- software development
- text mining
- knowledge representation
- natural language
- bayesian networks
- artificial intelligence