# Elucidating the Role of Feature Normalization in IJEPA

arXiv: https://arxiv.org/abs/2508.02829 · HTML: https://mlsanity.com/paper/2508.02829/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in General knowledge distillation methods, one of 12 directions in Knowledge distillation.
- Builds on 13 of its 16 references that are in the map.
- First posted on arXiv 4 Aug 2025.

## Placed in

- Knowledge distillation / [General knowledge distillation methods](https://mlsanity.com/distill/general-knowledge-distillation-methods/)

## Builds on (most cited first, 13 of 13)

- [A Simple Framework for Contrastive Learning of Visual Representations](https://mlsanity.com/paper/2002.05709/) (2020)
- [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)
- [Bootstrap your own latent: A new approach to self-supervised Learning](https://mlsanity.com/paper/2006.07733/) (2020)
- [Vision Transformers for Dense Prediction](https://mlsanity.com/paper/2103.13413/) (2021)
- [Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfer](https://mlsanity.com/paper/1907.01341/) (2019)
- [Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data](https://mlsanity.com/paper/2401.10891/) (2024)
- [Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Think](https://mlsanity.com/paper/2410.06940/) (2024)
- [Perception Encoder: The best visual embeddings are not at the output of the network](https://mlsanity.com/paper/2504.13181/) (2025)
- [RankMe: Assessing the downstream performance of pretrained self-supervised representations by their rank](https://mlsanity.com/paper/2210.02885/) (2022)
- [Deconstructing Denoising Diffusion Models for Self-Supervised Learning](https://mlsanity.com/paper/2401.14404/) (2024)
- [LiDAR: Sensing Linear Probing Performance in Joint Embedding SSL Architectures](https://mlsanity.com/paper/2312.04000/) (2023)
- [DiffMoE: Dynamic Token Selection for Scalable Diffusion Transformers](https://mlsanity.com/paper/2503.14487/) (2025)

## Abstract (arXiv)

In the standard image joint embedding predictive architecture (IJEPA), features at the output of the teacher encoder are layer normalized (LN) before serving as a distillation target for the student encoder and predictor. We propose that this feature normalization disrupts the natural energy hierarchy of visual tokens, where high-energy tokens (those with larger L2 norms) encode semantically important image regions. LN forces all features to have identical L2 norms, effectively equalizing their energies and preventing the model from prioritizing semantically rich regions. We find that IJEPA models trained with feature LN exhibit loss maps with significant checkerboard-like artifacts. We propose that feature LN be replaced with a DynTanh activation as the latter better preserves token energies and allows high-energy tokens to greater contribute to the prediction loss. We show that IJEPA trained with feature DynTanh exhibits a longer-tailed loss distribution and fixes the checkerboard artifacts in the loss map. Our empirical results show that our simple modification improves ImageNet linear probe accuracy from 38% to 42.7% for ViT-Small and reduces RMSE by 0.08 on NYU Depth V2 monocular depth estimation. These results suggest that preserving natural token energies is crucial for effective self-supervised visual representation learning.