Choosing Where and How to Moderate: End-to-End Trade-offs in Filter Placement and Response Rewriting
Content-moderation classifiers are usually evaluated in isolation, but deployment requires choosing where to intervene and what follows a flag. We evaluate these choices using two end-to-end customer-outcome metrics rather than component accuracy: Usefulness, the fraction of turns with a shown, non-harmful, relevant response, and Harmful Exposure, the fraction with a shown harmful response. Latency and error rates are diagnostics. We compare Input only, Response only, and Input + response hard blocking on a human-labelled product benchmark and public ToxicChat evaluation. At the evaluated operating points, Response only achieves the highest filter-only Usefulness in both settings, while I...
| Authors | Mengya Hu, Susie Park, Suzana Ilic et al. |
|---|---|
| Primary category | cs.CL |
| Published | 28 July 2026 |
| Last updated at source | 28 July 2026 |
Silytics shows public metadata from arXiv and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.