# Not All Diffusion Model Activations Have Been Evaluated as Discriminative Features

arXiv: https://arxiv.org/abs/2410.03558 · HTML: https://mlsanity.com/paper/2410.03558/

## Key facts

- Cited by 25 papers in the map (data through 14 Oct 2026).
- Placed in Image generation, text-to-image and editing, one of 14 directions in Diffusion models.
- Builds on 50 of its 67 references that are in the map.
- First posted on arXiv 4 Oct 2024.

## Placed in

- Diffusion models / [Image generation, text-to-image and editing](https://mlsanity.com/diffusion/image-generation-text-to-image-and-editing/)

## Builds on (most cited first, 30 of 50)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2006.11239/) (2020)
- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [High-Resolution Image Synthesis with Latent Diffusion Models](https://mlsanity.com/paper/2112.10752/) (2021)
- [Score-Based Generative Modeling through Stochastic Differential Equations](https://mlsanity.com/paper/2011.13456/) (2020)
- [Diffusion Models Beat GANs on Image Synthesis](https://mlsanity.com/paper/2105.05233/) (2021)
- [Masked Autoencoders Are Scalable Vision Learners](https://mlsanity.com/paper/2111.06377/) (2021)
- [Emerging Properties in Self-Supervised Vision Transformers](https://mlsanity.com/paper/2104.14294/) (2021)
- [Scalable Diffusion Models with Transformers](https://mlsanity.com/paper/2212.09748/) (2022)
- [Flow Matching for Generative Modeling](https://mlsanity.com/paper/2210.02747/) (2022)
- [The Cityscapes Dataset for Semantic Urban Scene Understanding](https://mlsanity.com/paper/1604.01685/) (2016)
- [SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis](https://mlsanity.com/paper/2307.01952/) (2023)
- [Unsupervised Learning of Visual Features by Contrasting Cluster Assignments](https://mlsanity.com/paper/2006.09882/) (2020)
- [LSUN: Construction of a Large-scale Image Dataset using Deep Learning with Humans in the Loop](https://mlsanity.com/paper/1506.03365/) (2015)
- [Transformer Feed-Forward Layers Are Key-Value Memories](https://mlsanity.com/paper/2012.14913/) (2020)
- [VideoCrafter2: Overcoming Data Limitations for High-Quality Video Diffusion Models](https://mlsanity.com/paper/2401.09047/) (2024)
- [Label-Efficient Semantic Segmentation with Diffusion Models](https://mlsanity.com/paper/2112.03126/) (2021)
- [Emergent Correspondence from Image Diffusion](https://mlsanity.com/paper/2306.03881/) (2023)
- [Open-Vocabulary Panoptic Segmentation with Text-to-Image Diffusion Models](https://mlsanity.com/paper/2303.04803/) (2023)
- [Road Extraction by Deep Residual U-Net](https://mlsanity.com/paper/1711.10684/) (2017)
- [Deep ViT Features as Dense Visual Descriptors](https://mlsanity.com/paper/2112.05814/) (2021)
- [Repurposing Diffusion-Based Image Generators for Monocular Depth Estimation](https://mlsanity.com/paper/2312.02145/) (2023)
- [Your Diffusion Model is Secretly a Zero-Shot Classifier](https://mlsanity.com/paper/2303.16203/) (2023)
- [A Tale of Two Features: Stable Diffusion Complements DINO for Zero-Shot Semantic Correspondence](https://mlsanity.com/paper/2305.15347/) (2023)
- [Unleashing Text-to-Image Diffusion Models for Visual Perception](https://mlsanity.com/paper/2303.02153/) (2023)
- [Diffusion Hyperfeatures: Searching Through Time and Space for Semantic Correspondence](https://mlsanity.com/paper/2305.14334/) (2023)
- [Dual Diffusion Implicit Bridges for Image-to-Image Translation](https://mlsanity.com/paper/2203.08382/) (2022)
- [MaskCLIP: Masked Self-Distillation Advances Contrastive Language-Image Pretraining](https://mlsanity.com/paper/2208.12262/) (2022)

## Built on by (most cited first, 25 of 25)

- [Emergent Temporal Correspondences from Video Diffusion Transformers](https://mlsanity.com/paper/2506.17220/) (2025)
- [Seg4Diff: Unveiling Open-Vocabulary Segmentation in Text-to-Image Diffusion Transformers](https://mlsanity.com/paper/2509.18096/) (2025)
- [Aligning Text to Image in Diffusion Models is Easier Than You Think](https://mlsanity.com/paper/2503.08250/) (2025)
- [MixBridge: Heterogeneous Image-to-Image Backdoor Attack through Mixture of Schr"odinger Bridges](https://mlsanity.com/paper/2505.08809/) (2025)
- [Guiding Registration with Emergent Similarity from Pre-Trained Diffusion Models](https://mlsanity.com/paper/2506.02419/) (2025)
- [StyleKeeper: Prevent Content Leakage using Negative Visual Query Guidance](https://mlsanity.com/paper/2510.06827/) (2025)
- [Segment Anyword: Mask Prompt Inversion for Open-Set Grounded Segmentation](https://mlsanity.com/paper/2505.17994/) (2025)
- [Appreciate the View: A Task-Aware Evaluation Framework for Novel View Synthesis](https://mlsanity.com/paper/2511.12675/) (2025)
- [Towards Multimodal Understanding via Stable Diffusion as a Task-Aware Feature Extractor](https://mlsanity.com/paper/2507.07106/) (2025)
- [Probing and Leveraging Video Diffusion Transformer Features for Robust Point Tracking](https://mlsanity.com/paper/2512.20606/) (2025)
- [Making Training-Free Diffusion Segmentors Scale with the Generative Power](https://mlsanity.com/paper/2603.06178/) (2026)
- [GenMask: Adapting DiT for Segmentation via Direct Mask Generation](https://mlsanity.com/paper/2603.23906/) (2026)
- [Guiding a Diffusion Model by Swapping Its Tokens](https://mlsanity.com/paper/2604.08048/) (2026)
- [Leveraging Prior Knowledge of Diffusion Model for Person Search](https://mlsanity.com/paper/2510.01841/) (2025)
- [BachVid: Training-Free Video Generation with Consistent Background and Character](https://mlsanity.com/paper/2510.21696/) (2025)
- [Self-transcendence: Is External Feature Guidance Indispensable for Accelerating Diffusion Transformer Training?](https://mlsanity.com/paper/2601.07773/) (2026)
- [Canonical Latent Representations in Conditional Diffusion Models](https://mlsanity.com/paper/2506.09955/) (2025)
- [Depth-Sequence Transformer (DST) for Segment-Specific ICA Calcification Mapping on Non-Contrast CT](https://mlsanity.com/paper/2507.08214/) (2025)
- [Exploring Conditions for Diffusion models in Robotic Control](https://mlsanity.com/paper/2510.15510/) (2025)
- [General and Efficient Steering of Diffusion Models](https://mlsanity.com/paper/2602.11395/) (2026)
- [BiGain: Unified Token Compression for Joint Generation and Classification](https://mlsanity.com/paper/2603.12240/) (2026)
- [Noise2Map: End-to-End Diffusion Model for Semantic Segmentation and Change Detection](https://mlsanity.com/paper/2604.27889/) (2026)
- [Diffuse2Seg: Diffusion Models Can Segment Anything Without Supervision](https://mlsanity.com/paper/2609.06491/) (2026)
- [Why DiT Models Underperform as Representation Learners without Long Skip Connections](https://mlsanity.com/paper/2610.13344/) (2026)
- [We Have Been Interpreting Diffusion Transformer Backbone Behaviors Wrong](https://mlsanity.com/paper/2610.14517/) (2026)

## Abstract (arXiv)

Diffusion models are initially designed for image generation. Recent research shows that the internal signals within their backbones, named activations, can also serve as dense features for various discriminative tasks such as semantic segmentation. Given numerous activations, selecting a small yet effective subset poses a fundamental problem. To this end, the early study of this field performs a large-scale quantitative comparison of the discriminative ability of the activations. However, we find that many potential activations have not been evaluated, such as the queries and keys used to compute attention scores. Moreover, recent advancements in diffusion architectures bring many new activations, such as those within embedded ViT modules. Both combined, activation selection remains unresolved but overlooked. To tackle this issue, this paper takes a further step with a much broader range of activations evaluated. Considering the significant increase in activations, a full-scale quantitative comparison is no longer operational. Instead, we seek to understand the properties of these activations, such that the activations that are clearly inferior can be filtered out in advance via simple qualitative evaluation. After careful analysis, we discover three properties universal among diffusion models, enabling this study to go beyond specific models. On top of this, we present effective feature selection solutions for several popular diffusion models. Finally, the experiments across multiple discriminative tasks validate the superiority of our method over the SOTA competitors. Our code is available at https://github.com/Darkbblue/generic-diffusion-feature.