Pinned Loading

1. aimagelab/safe-clip aimagelab/safe-clip Public

   [ECCV 2024] Official implementation of Safe-CLIP, a safety-aligned vision-language model for mitigating NSFW content in multimodal retrieval and generation.

   Python 68

2. aimagelab/HySAC aimagelab/HySAC Public

   [CVPR 2025 Highlight] Official implementation of HySAC, a hyperbolic safety-aware vision-language model for safer multimodal retrieval and content moderation.

   Python 31 1

3. aimagelab/cross-model-safety-representations aimagelab/cross-model-safety-representations Public

   [Under Review] Official implementation of cross-model safety steering, a framework that transfers safety directions from LLMs to heterogeneous image and video generators.

   7

4. aimagelab/ShieldCLIP aimagelab/ShieldCLIP Public

   Official repository for ShieldCLIP: selective safety alignment for harmful content mitigation in multimodal foundation models. Code, trained models, and dataset access instructions coming soon.

   2

5. aimagelab/CounterVid aimagelab/CounterVid Public

   [EMNLP 2026] Official implementation of CounterVid, a counterfactual video generation framework for mitigating action and temporal hallucinations in video-language models.

   Python 4