LLM hidden states from the reading phase encode domain-specific signals that can route queries to better models, boosting average accuracy by 12.3% over a single fine-tuned model.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Exploring How LLMs Capture and Represent Domain-Specific Knowledge
LLM hidden states from the reading phase encode domain-specific signals that can route queries to better models, boosting average accuracy by 12.3% over a single fine-tuned model.