# Safe-CLIP: Removing NSFW Concepts from Vision-and-Language Models

arXiv: https://arxiv.org/abs/2311.16254 · HTML: https://mlsanity.com/paper/2311.16254/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- Builds on 47 of its 63 references that are in the map.
- First posted on arXiv 27 Nov 2023.

## Builds on (most cited first, 30 of 47)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [High-Resolution Image Synthesis with Latent Diffusion Models](https://mlsanity.com/paper/2112.10752/) (2021)
- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [Representation Learning with Contrastive Predictive Coding](https://mlsanity.com/paper/1807.03748/) (2018)
- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://mlsanity.com/paper/1910.01108/) (2019)
- [LAION-5B: An open large-scale dataset for training next generation image-text models](https://mlsanity.com/paper/2210.08402/) (2022)
- [Deep reinforcement learning from human preferences](https://mlsanity.com/paper/1706.03741/) (2017)
- [GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models](https://mlsanity.com/paper/2112.10741/) (2021)
- [QLoRA: Efficient Finetuning of Quantized LLMs](https://mlsanity.com/paper/2305.14314/) (2023)
- [Self-Instruct: Aligning Language Models with Self-Generated Instructions](https://mlsanity.com/paper/2212.10560/) (2022)
- [Evaluating Object Hallucination in Large Vision-Language Models](https://mlsanity.com/paper/2305.10355/) (2023)
- [MME: A Comprehensive Evaluation Benchmark for Multimodal Large Language Models](https://mlsanity.com/paper/2306.13394/) (2023)
- [LAION-400M: Open Dataset of CLIP-Filtered 400 Million Image-Text Pairs](https://mlsanity.com/paper/2111.02114/) (2021)
- [LIMA: Less Is More for Alignment](https://mlsanity.com/paper/2305.11206/) (2023)
- [Deep Visual-Semantic Alignments for Generating Image Descriptions](https://mlsanity.com/paper/1412.2306/) (2014)
- [A Diagram Is Worth A Dozen Images](https://mlsanity.com/paper/1603.07396/) (2016)
- [LLaMA-Adapter: Efficient Fine-tuning of Language Models with Zero-init Attention](https://mlsanity.com/paper/2303.16199/) (2023)
- [DataComp: In search of the next generation of multimodal datasets](https://mlsanity.com/paper/2304.14108/) (2023)
- [Eternal Sunshine of the Spotless Net: Selective Forgetting in Deep Networks](https://mlsanity.com/paper/1911.04933/) (2019)
- [LLaMA-Adapter V2: Parameter-Efficient Visual Instruction Model](https://mlsanity.com/paper/2304.15010/) (2023)
- [Erasing Concepts from Diffusion Models](https://mlsanity.com/paper/2303.07345/) (2023)

## Built on by (most cited first, 12 of 12)

- [The Revolution of Multimodal Large Language Models: A Survey](https://mlsanity.com/paper/2402.12451/) (2024)
- [Latent Guard: a Safety Framework for Text-to-image Generation](https://mlsanity.com/paper/2404.08031/) (2024)
- [The Stronger the Diffusion Model, the Easier the Backdoor: Data Poisoning to Induce Copyright Breaches Without Adjusting Finetuning Pipeline](https://mlsanity.com/paper/2401.04136/) (2024)
- [AlignGuard: Scalable Safety Alignment for Text-to-Image Generation](https://mlsanity.com/paper/2412.10493/) (2024)
- [Is Retain Set All You Need in Machine Unlearning? Restoring Performance of Unlearned Models with Out-Of-Distribution Images](https://mlsanity.com/paper/2404.12922/) (2024)
- [Survey of Video Diffusion Models: Foundations, Implementations, and Applications](https://mlsanity.com/paper/2504.16081/) (2025)
- [DUCK: Distance-based Unlearning via Centroid Kinematics](https://mlsanity.com/paper/2312.02052/) (2023)
- [SteerDiff: Steering towards Safe Text-to-Image Diffusion Models](https://mlsanity.com/paper/2410.02710/) (2024)
- [A Comprehensive Survey on Concept Erasure in Text-to-Image Diffusion Models](https://mlsanity.com/paper/2502.14896/) (2025)
- [CRCE: Coreference-Retention Concept Erasure in Text-to-Image Diffusion Models](https://mlsanity.com/paper/2503.14232/) (2025)
- [Safety Without Semantic Disruptions: Editing-free Safe Image Generation via Context-preserving Dual Latent Reconstruction](https://mlsanity.com/paper/2411.13982/) (2024)
- [Harmful Content Generation in Text-to-Image Models: Capabilities and Moderation Limitations](https://mlsanity.com/paper/2610.06503/) (2026)

## Abstract (arXiv)

Large-scale vision-and-language models, such as CLIP, are typically trained on web-scale data, which can introduce inappropriate content and lead to the development of unsafe and biased behavior. This, in turn, hampers their applicability in sensitive and trustworthy contexts and could raise significant concerns in their adoption. Our research introduces a novel approach to enhancing the safety of vision-and-language models by diminishing their sensitivity to NSFW (not safe for work) inputs. In particular, our methodology seeks to sever "toxic" linguistic and visual concepts, unlearning the linkage between unsafe linguistic or visual items and unsafe regions of the embedding space. We show how this can be done by fine-tuning a CLIP model on synthetic data obtained from a large language model trained to convert between safe and unsafe sentences, and a text-to-image generator. We conduct extensive experiments on the resulting embedding space for cross-modal retrieval, text-to-image, and image-to-text generation, where we show that our model can be remarkably employed with pre-trained generative models. Our source code and trained models are available at: https://github.com/aimagelab/safe-clip.