OpenAI releases open-source reasoning model GPT-OSS Safeguard for safety classification tasks
GPT-OSS-Safeguard allows users to classify content using custom policies. The model interprets these policies to classify messages, replies, and conversations. It is suitable for scenarios where potential harm arises, or for policies that are constantly evolving and require rapid adjustments. GPT-OSS-Safeguard is licensed under Apache 2.0.…