Llama Guard 3 Vision: Safeguarding Human-AI Image Understanding Conversations
This addresses content moderation for multimodal AI conversations, though it is incremental as it extends previous text-only versions to support images.
The paper tackles the problem of safeguarding human-AI conversations involving image understanding by introducing Llama Guard 3 Vision, a multimodal LLM-based safeguard that detects harmful multimodal prompts and text responses, demonstrating strong performance on internal benchmarks.
We introduce Llama Guard 3 Vision, a multimodal LLM-based safeguard for human-AI conversations that involves image understanding: it can be used to safeguard content for both multimodal LLM inputs (prompt classification) and outputs (response classification). Unlike the previous text-only Llama Guard versions (Inan et al., 2023; Llama Team, 2024b,a), it is specifically designed to support image reasoning use cases and is optimized to detect harmful multimodal (text and image) prompts and text responses to these prompts. Llama Guard 3 Vision is fine-tuned on Llama 3.2-Vision and demonstrates strong performance on the internal benchmarks using the MLCommons taxonomy. We also test its robustness against adversarial attacks. We believe that Llama Guard 3 Vision serves as a good starting point to build more capable and robust content moderation tools for human-AI conversation with multimodal capabilities.