ShieldGemma: Generative AI Content Moderation Based on Gemma
This provides a resource for developers to improve content moderation in LLMs, addressing safety risks like hate speech and harassment, but it is incremental as it builds on existing models like Gemma2.
The authors tackled the problem of content moderation for generative AI by developing ShieldGemma, a suite of LLM-based safety models built on Gemma2, which achieved superior performance with gains of +10.8% AU-PRC over Llama Guard and +4.3% over WildCard on benchmarks.
We present ShieldGemma, a comprehensive suite of LLM-based safety content moderation models built upon Gemma2. These models provide robust, state-of-the-art predictions of safety risks across key harm types (sexually explicit, dangerous content, harassment, hate speech) in both user input and LLM-generated output. By evaluating on both public and internal benchmarks, we demonstrate superior performance compared to existing models, such as Llama Guard (+10.8\% AU-PRC on public benchmarks) and WildCard (+4.3\%). Additionally, we present a novel LLM-based data curation pipeline, adaptable to a variety of safety-related tasks and beyond. We have shown strong generalization performance for model trained mainly on synthetic data. By releasing ShieldGemma, we provide a valuable resource to the research community, advancing LLM safety and enabling the creation of more effective content moderation solutions for developers.