A survey that catalogs threat models, detection approaches, and mitigation strategies for toxicity in multilingual LLMs while identifying challenges such as uneven language coverage and culturally variable harm definitions.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
A Survey of Toxicity Detection and Mitigation Strategies for Multilingual Language Models
A survey that catalogs threat models, detection approaches, and mitigation strategies for toxicity in multilingual LLMs while identifying challenges such as uneven language coverage and culturally variable harm definitions.