arXiv:2608. 14922v1 Announce Type: cross Abstract: Mechanistic interpretability has recently expanded to Vision Transformers (ViTs), with Sparse Autoencoders (SAEs) increasingly used as post-hoc tools to decompose internal representations into sparse and more interpretable features.
arXiv:2606. 06664v1 Announce Type: cross Abstract: Despite high accuracy, Vision Transformer (ViT) predictions can be driven by spurious cues, raising the need to understand their inner workings before safe deployment.
arXiv:2607. 08605v1 Announce Type: cross Abstract: Sparse autoencoders (SAEs) have emerged as a promising technique for mechanistic interpretability by learning a set of sparse latent features in large models, each of which encodes a distinct concept.
arXiv:2606. 24716v1 Announce Type: cross Abstract: Sparse autoencoders (SAEs) are increasingly used to extract interpretable concepts from vision and vision language models, yet existing evaluation methods largely rely on proxy metrics or qualitative inspection rather than measuring semantic correspondence.
arXiv:2607. 00089v1 Announce Type: new Abstract: Mechanistic interpretability has produced a rich inventory of component-level analyses that characterise what neural-network components encode and how they interact.
ViTAMINS is a method that incorporates synthetic hard negatives into unsupervised vision transformer pretraining to enhance representation quality. The approach is evaluated on ImageNet and a range of downstream tasks—including transfer learning, image retrieval, copy detection, and image/video segmentation—showing significant performance gains. The synthetic negatives also lead to emergent properties, such as representations that encode explicit semantic information and act as strong classifiers, improving over baselines by up to 11.3%.
Spikformer V2 introduces a Spiking Self‑Attention (SSA) mechanism that removes softmax and uses spike‑based Query, Key, and Value to capture sparse visual features efficiently. It also adds a Spiking Convolutional Stem (SCS) and employs self‑supervised learning (masking and reconstruction) to pre‑train the model before fine‑tuning on ImageNet. The result is the first spiking neural network to surpass 80 % accuracy on ImageNet, achieving 81.10 % with a 172 M‑parameter, 16‑layer model in just one time step.
arXiv:2607. 10391v1 Announce Type: cross Abstract: Despite exposing rich intermediate representations, Vision Transformers (ViTs) are almost exclusively utilized as black-box feature extractors, where only the last layer is considered for downstream tasks.
arXiv:2607. 07316v1 Announce Type: new Abstract: This article offers a comprehensive overview of mechanistic interpretability, an emerging field that seeks to reverse-engineer the internal algorithms of modern neural networks.
arXiv:2605.05556v2 Announce Type: replace Abstract: Artificial neural networks trained on visual tasks develop internal representations resembling those of the primate visual system, a discovery that...
arXiv:2511. 09432v2 Announce Type: replace Abstract: Machine learning (ML) models achieve remarkable performance but remain hard to interpret due to their scale and complexity.
arXiv:2607.09086v2 Announce Type: replace Abstract: We present Subtoken Vision Transformer (SubViT), a selective image tokenization method for fine-grained visual recognition. Standard Vision Transfo...
arXiv:2609.24379v1 Announce Type: cross Abstract: Mechanistic interpretability of vision transformers seeks to decompose model computation into human-readable units, but learned representations entan...