PolicyLM-1.7B Applies Plain-English Moderation Policies in Under 50ms

Musubi announced PolicyLM-1.7B, a decision model designed to enforce a content policy written in plain English in under 50 milliseconds. The company says it applies complex policies without special training and needs no retraining when the policy changes, at cost and speed similar to classifiers.