Fine-tuning a 1B LLaMA model on a synthetic emotional-safety benchmark matches or beats 70B few-shot performance and a BERT baseline on three high-data categories, using under 2GB VRAM.
Transformers reinforcement learning (trl),
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Scaling behavior of large language models in emotional safety classification across sizes and tasks
Fine-tuning a 1B LLaMA model on a synthetic emotional-safety benchmark matches or beats 70B few-shot performance and a BERT baseline on three high-data categories, using under 2GB VRAM.