Open Source · MarkTechPost ·
Fastino Labs Open-Sources GLiGuard: A 300M Parameter Safety Moderation Model That Matches or Exceeds Accuracy of Models 23–90x Its Size
Fastino Labs has open-sourced GLiGuard, a 300M-parameter safety moderation model that handles prompt safety, jailbreak detection, harm classification, and refusal detection in one forward pass. The model uses an encoder architecture and is reported to match or exceed larger models on nine safety benchmarks while offeri