UDRIL routes only high-uncertainty user queries to a fine-tuned LLM, improving out-of-scope intent detection while keeping most predictions on a cheap classifier.
Title resolution pending
1 Pith paper cite this work, alongside 12 external citations. Polarity classification is still indexing.
1
Pith paper citing it
12
external citations · OpenAlex
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Efficient Out-of-Scope Detection in Dialogue Systems via Uncertainty-Driven LLM Routing
UDRIL routes only high-uncertainty user queries to a fine-tuned LLM, improving out-of-scope intent detection while keeping most predictions on a cheap classifier.