A blind spot for large language models: Supradiegetic linguistic information.
Julia Witte ZimmermanDenis HudonKathryn CramerJonathan St-OngeMikaela Irene D. FudoligMilo Z. TrujilloChristopher M. DanforthPeter Sheridan DoddsPublished in: CoRR (2023)
Keyphrases
- language model
- linguistic information
- multiword
- language modeling
- n gram
- part of speech
- linguistic features
- structural information
- information retrieval
- query expansion
- retrieval model
- semantic information
- probabilistic model
- test collection
- document retrieval
- translation model
- context sensitive
- vector space model
- relevance model
- pseudo relevance feedback
- query terms
- domain knowledge