Fine-tuning small models with SFT plus DPO on empathy-improved responses from LLMs yields large judged win-rate gains over base small models.
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
-
Distilling Empathy from Large Language Models
Fine-tuning small models with SFT plus DPO on empathy-improved responses from LLMs yields large judged win-rate gains over base small models.