# Gate-and-Merge: Zero-shot Compositional Personalization of Vision Language Models

arXiv: https://arxiv.org/abs/2605.08702 · HTML: https://mlsanity.com/paper/2605.08702/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Prompt learning and adaptation of CLIP, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 22 of its 24 references that are in the map.
- First posted on arXiv 9 May 2026.

## Placed in

- Multimodal LLMs / vision-language models / [Prompt learning and adaptation of CLIP](https://mlsanity.com/mllm/prompt-learning-and-adaptation-of-clip/)

## Builds on (most cited first, 22 of 22)

- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models](https://mlsanity.com/paper/2301.12597/) (2023)
- [DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation](https://mlsanity.com/paper/2208.12242/) (2022)
- [Conditional Prompt Learning for Vision-Language Models](https://mlsanity.com/paper/2203.05557/) (2022)
- [Conceptual 12M: Pushing Web-Scale Image-Text Pre-Training To Recognize Long-Tail Visual Concepts](https://mlsanity.com/paper/2102.08981/) (2021)
- [TIES-Merging: Resolving Interference When Merging Models](https://mlsanity.com/paper/2306.01708/) (2023)
- [Language Models are Super Mario: Absorbing Abilities from Homologous Models as a Free Lunch](https://mlsanity.com/paper/2311.03099/) (2023)
- [VisionLLM: Large Language Model is also an Open-Ended Decoder for Vision-Centric Tasks](https://mlsanity.com/paper/2305.11175/) (2023)
- [Tip-Adapter: Training-free Adaption of CLIP for Few-shot Classification](https://mlsanity.com/paper/2207.09519/) (2022)
- [HydraLoRA: An Asymmetric LoRA Architecture for Efficient Fine-Tuning](https://mlsanity.com/paper/2404.19245/) (2024)
- [Orthogonal Adaptation for Modular Customization of Diffusion Models](https://mlsanity.com/paper/2312.02432/) (2023)
- [Yo'LLaVA: Your Personalized Language and Vision Assistant](https://mlsanity.com/paper/2406.09400/) (2024)
- [MyVLM: Personalizing VLMs for User-Specific Queries](https://mlsanity.com/paper/2403.14599/) (2024)
- [MC-LLaVA: Multi-Concept Personalized Vision-Language Model](https://mlsanity.com/paper/2411.11706/) (2024)
- [JeDi: Joint-Image Diffusion Models for Finetuning-Free Personalized Text-to-Image Generation](https://mlsanity.com/paper/2407.06187/) (2024)
- [LoRA Soups: Merging LoRAs for Practical Skill Composition Tasks](https://mlsanity.com/paper/2410.13025/) (2024)
- [Universality and Limitations of Prompt Tuning](https://mlsanity.com/paper/2305.18787/) (2023)
- [RAP: Retrieval-Augmented Personalization for Multimodal Large Language Models](https://mlsanity.com/paper/2410.13360/) (2024)
- [Personalized Image Generation with Deep Generative Models: A Decade Survey](https://mlsanity.com/paper/2502.13081/) (2025)
- [Personalized Image Generation with Large Multimodal Models](https://mlsanity.com/paper/2410.14170/) (2024)
- [MultLFG: Training-free Multi-LoRA composition using Frequency-domain Guidance](https://mlsanity.com/paper/2505.20525/) (2025)
- [Compositional Subspace Representation Fine-tuning for Adaptive Large Language Models](https://mlsanity.com/paper/2503.10617/) (2025)

## Abstract (arXiv)

This paper tackles compositional personalization of vision-language models (VLMs). In this problem, multiple user-defined concepts must be recognized or described jointly at test time. We introduce Gate-and-Merge, a zero-shot framework that enables compositional personalization without the need for co-occurrence training. During personalization, each concept is learned independently as a lightweight LoRA adapter, paired with a concept token. The base model remains unchanged and concepts are kept disentangled. At inference, we enable composition by merging concept-specific LoRA updates directly in weight space. To suppress irrelevant activations and prevent interference, a gating mechanism is employed to estimate textual and visual cues and select only the modules that contribute to the prediction. We further stabilize composition by combining only the most meaningful and mutually consistent updates, helping preserve each concept's identity. Our quantitative and qualitative analyses show consistent gains in performance across multiple personalization tasks in both single-concept and compositional settings.