Blog News, updates, and engineering notes

Methodology

The Brutal Truth About Our Content Moderation API: We Tested Against Industry Benchmarks and Here's What Broke

We ran Sentiqra against HateCheck and Jigsaw Civil Comments — the two most unforgiving public benchmarks in content moderation. Our F1 was 62.7%. Here is everything we learned, including where we lost.

Methodology

Building a Multilingual Moderation Pipeline: Lessons from 14 Languages

Calibrating a detection system across 14 languages is not a translation problem. It is a cultural, linguistic, and engineering problem. Here is what we learned.