# DriftScope: Measuring The Hidden Effects of Diffusion Model Adaptation

arXiv: https://arxiv.org/abs/2607.00183 · HTML: https://mlsanity.com/paper/2607.00183/

## Key facts

- Cited by 1 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 30 of its 37 references that are in the map.
- First posted on arXiv 30 Jun 2026.

## 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 30)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2006.11239/) (2020)
- [The Unreasonable Effectiveness of Deep Features as a Perceptual Metric](https://mlsanity.com/paper/1801.03924/) (2018)
- [Score-Based Generative Modeling through Stochastic Differential Equations](https://mlsanity.com/paper/2011.13456/) (2020)
- [Deep Unsupervised Learning using Nonequilibrium Thermodynamics](https://mlsanity.com/paper/1503.03585/) (2015)
- [DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation](https://mlsanity.com/paper/2208.12242/) (2022)
- [An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversion](https://mlsanity.com/paper/2208.01618/) (2022)
- [Multi-Concept Customization of Text-to-Image Diffusion](https://mlsanity.com/paper/2212.04488/) (2022)
- [Erasing Concepts from Diffusion Models](https://mlsanity.com/paper/2303.07345/) (2023)
- [Class-incremental learning: survey and performance evaluation on image classification](https://mlsanity.com/paper/2010.15277/) (2020)
- [SVDiff: Compact Parameter Space for Diffusion Fine-Tuning](https://mlsanity.com/paper/2303.11305/) (2023)
- [MACE: Mass Concept Erasure in Diffusion Models](https://mlsanity.com/paper/2403.06135/) (2024)
- [Ring-A-Bell! How Reliable are Concept Removal Methods for Diffusion Models?](https://mlsanity.com/paper/2310.10012/) (2023)
- [One-Dimensional Adapter to Rule Them All: Concepts, Diffusion Models and Erasing Applications](https://mlsanity.com/paper/2312.16145/) (2023)
- [Dynamical Regimes of Diffusion Models](https://mlsanity.com/paper/2402.18491/) (2024)
- [Continual Diffusion: Continual Customization of Text-to-Image Diffusion with C-LoRA](https://mlsanity.com/paper/2304.06027/) (2023)
- [Parameter-Efficient Orthogonal Finetuning via Butterfly Factorization](https://mlsanity.com/paper/2311.06243/) (2023)
- [Erasing Undesirable Influence in Diffusion Models](https://mlsanity.com/paper/2401.05779/) (2024)
- [Scissorhands: Scrub Data Influence via Connection Sensitivity in Networks](https://mlsanity.com/paper/2401.06187/) (2024)
- [Mass-Producing Failures of Multimodal Systems with Language Models](https://mlsanity.com/paper/2306.12105/) (2023)
- [Q-Eval-100K: Evaluating Visual Quality and Alignment Level for Text-to-Vision Content](https://mlsanity.com/paper/2503.02357/) (2025)
- [Discovering Failure Modes of Text-guided Diffusion Models via Adversarial Search](https://mlsanity.com/paper/2306.00974/) (2023)
- [Exploring Continual Learning of Diffusion Models](https://mlsanity.com/paper/2303.15342/) (2023)
- [Multiscale Sliced Wasserstein Distances as Perceptual Color Difference Measures](https://mlsanity.com/paper/2407.10181/) (2024)
- [A Simple and Efficient Baseline for Zero-Shot Generative Classification](https://mlsanity.com/paper/2412.12594/) (2024)
- [Uncovering Conceptual Blindspots in Generative Image Models Using Sparse Autoencoders](https://mlsanity.com/paper/2506.19708/) (2025)
- [DEXTER: Diffusion-Guided EXplanations with TExtual Reasoning for Vision Models](https://mlsanity.com/paper/2510.14741/) (2025)
- [Assessing Open-world Forgetting in Generative Image Model Customization](https://mlsanity.com/paper/2410.14159/) (2024)
- [Discovering Divergent Representations between Text-to-Image Models](https://mlsanity.com/paper/2509.08940/) (2025)
- [Forget Less by Learning Together through Concept Consolidation](https://mlsanity.com/paper/2601.01963/) (2026)

## Built on by (most cited first, 1 of 1)

- [Discovering Undisclosed Concepts in LoRA Adapters for Diffusion-Based AIGC Models](https://mlsanity.com/paper/2610.18990/) (2026)

## Abstract (arXiv)

Adapting pre-trained text-to-image diffusion models, whether to learn new visual concepts or erase unwanted ones, is routinely evaluated on its intended effects alone. We argue this framing is incomplete. Through sparse autoencoder analysis and zero-shot classification, we demonstrate that adaptation systematically damages semantically unrelated concepts in ways that aggregate metrics structurally cannot surface: when damage is severe enough for FID and KID to respond, the model is already nearly unusable; when the model remains functional, FID and KID stay flat while specific classes silently suffer worst-case zero-shot accuracy drops of up to 18.9 points and concept-level distributions shift dramatically. This pattern appears at both ends of the adaptation spectrum (concept customization and concept unlearning), suggesting it is a systematic consequence of weight-level modification rather than an artifact of any particular method. To surface this hidden drift before deployment, we introduce DriftScope, a prompt-level diagnostic tool that takes any two model checkpoints and returns a ranked list of tokens whose visual concepts have shifted most between them. DriftScope optimizes a soft prompt to attribute drift at the token level without requiring access to real data or model internals. The result is an interpretable, concept-level audit that aggregate evaluation cannot provide.