We’ve created activation atlases (in collaboration with Google researchers), a new technique for visualizing what interactions between neurons can represent. As AI systems are deployed in increasingly sensitive contexts, having a better understanding of their internal decision-making processes will let us identify weaknesses and investigate failures.
arXiv:2507. 10005v2 Announce Type: replace Abstract: In recent years, graph-based machine learning techniques, such as reinforcement learning and graph neural networks, have garnered significant attention.
We’re introducing OpenAI Microscope, a collection of visualizations of every significant layer and neuron of eight vision “model organisms” which are often studied in interpretability. Microscope makes it easier to analyze the features that form inside these neural networks, and we hope it will help the research community as we move towards understanding these complicated systems.
arXiv:2608.28637v1 Announce Type: new Abstract: Autonomous scientific discovery systems can generate large numbers of research ideas, experiments, and manuscripts with minimal human intervention. As...
The paper introduces a new graphical notation, adapted from Penrose tensor notation, to design and represent interpretable AI architectures. Unlike symbolic equations or probabilistic models, this notation provides a global view of an architecture while directly mapping to PyTorch einsum code. The authors demonstrate its use on several interpretable models and on the Steerling-8B language model, revealing structural insights and enabling concise code generation.
The paper introduces a graphical notation, adapted from Penrose tensor notation, to design and represent interpretable AI architectures. This notation provides a global view of an architecture and maps directly onto PyTorch einsum code, enabling clear depiction of tensor manipulations. The authors apply the notation to several interpretable models—concept bottlenecks, sparse probes, prototype networks, neural additive models, and mixtures of linear models—and use it to diagram the key components of the Steerling-8B language model, revealing its residual structure and allowing a concise 33‑line PyTorch implementation.
The survey titled "When Vision Meets Graphs: A Survey on Graph Reasoning and Learning" reviews how visual depictions of graphs can be used as inputs for graph reasoning and learning. It highlights that while Graph Neural Networks dominate graph machine learning, most pipelines ignore the visual form of graphs, despite scientists routinely interpreting graphs visually. The paper organizes existing work into three threads—vision for graph reasoning, vision for graph learning, and scientific graphs—aiming to clarify current capabilities and chart a path toward foundation models that perceive and reason about graphs like scientists do.
arXiv:2607. 28989v1 Announce Type: new Abstract: Many learning problems require representations that reconcile direct input, nearby structure, and broader context.
arXiv:2502. 00684v2 Announce Type: replace-cross Abstract: Deep reinforcement learning (DRL) has successfully addressed many complex control problems.
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:2606. 15767v1 Announce Type: cross Abstract: Understanding when and why deep neural networks are uncertain is crucial for deploying reliable machine learning systems in safety-critical domains.
arXiv:2509. 10650v4 Announce Type: replace-cross Abstract: Effective analysis in neuroscience benefits significantly from robust conceptual frameworks.
Google DeepMind is transforming the mouse pointer into a context-aware AI partner. Move beyond the friction of traditional prompting with intuitive AI collaboration in Chrome and beyond.
