Composing a general Japanese instruction model with a finance-specialized model via CALM cross-attention improves Japanese financial benchmark scores beyond LoRA, even when trained on a different finance dataset.
Universal language model fine-tuning for text classification,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Enhancing Financial Domain Adaptation of Language Models via Model Augmentation
Composing a general Japanese instruction model with a finance-specialized model via CALM cross-attention improves Japanese financial benchmark scores beyond LoRA, even when trained on a different finance dataset.