A two-stage SFT plus GRPO framework with a composite toxicity and semantic-similarity reward achieves state-of-the-art detoxification performance on ParaDetox using only 20% of annotated data.
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
-
Text Detoxification: Data Efficiency, Semantic Preservation and Model Generalization
A two-stage SFT plus GRPO framework with a composite toxicity and semantic-similarity reward achieves state-of-the-art detoxification performance on ParaDetox using only 20% of annotated data.