Moderation systems increasingly rely on external toxicity tools, but those tools are unreliable under code-mixing, transliteration, slang, and language mismatch. We study the conditional reliability of toxicity priors in Indian multilingual and code-mixed short text: English toxicity, Indic abuse, and rule-based severity cues can be useful evidence, but only in certain linguistic and abuse-severity contexts. We propose ToxGate, a trust-fusion head that conditions each auxiliary signal on the encoder representation before adding it to the prediction state. Across three short-text abuse datasets, four transformer encoders, and five seeds per setting, ToxGate improves over matched plain encoders in 10 of 12 in-domain settings and 7 of 8 transfer settings. The largest and most interpretable gains occur in high-risk moderation slices, including explicit slurs, violent threats, and cross-dataset transfer. The broader lesson is that moderation systems should treat external toxicity tools and priors as conditional evidence rather than fixed features or ground truth. In focused ablations, source-specific gating gives the strongest results in transfer, severe-abuse slices, and high-risk triage.
Blogger's Review: The introduction of ToxGate offers a fresh perspective on toxicity detection, particularly in multilingual contexts, emphasizing the importance of conditional reliability. This method not only enhances detection accuracy but also points the way for future developments in moderation systems. Its performance in high-risk scenarios is particularly encouraging and warrants further research and application.