A LoRA adapter trained on unstructured text and merged into an instruct model transfers style with a modest instruction-following loss, outperforming few-shot prompting on content fidelity.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
StyleAdaptedLM: Enhancing Instruction Following Models with Efficient Stylistic Transfer
A LoRA adapter trained on unstructured text and merged into an instruct model transfers style with a modest instruction-following loss, outperforming few-shot prompting on content fidelity.