Do Not (Always) Look Right: Investigating the Capabilities of Decoder-Based Large Language Models for Sequence Labeling.
David DukicJan SnajderPublished in: CoRR (2024)
Keyphrases
- language model
- sequence labeling
- conditional random fields
- probabilistic model
- language modeling
- dependency parsing
- retrieval model
- structured prediction
- n gram
- document retrieval
- named entity recognition
- query expansion
- test collection
- information retrieval
- smoothing methods
- language models for information retrieval
- markov models
- generative model
- context sensitive
- crf model
- graphical models
- training data
- pseudo relevance feedback
- maximum likelihood
- pairwise