Jacobian matrix
PulseAugur coverage of Jacobian matrix — every cluster mentioning Jacobian matrix across labs, papers, and developer communities, ranked by signal.
6 day(s) with sentiment data
-
Robotics research introduces outcome-based representation learning
Researchers have developed a novel method for learning manipulation-sufficient representations in robotics, focusing on action outcomes rather than dense geometric states. This approach utilizes an action-conditioned ou…
-
New Jacobian Lens Tool Tests if AI Models Use Internal Signals
Researchers have developed a new interpretability tool called the Jacobian lens, designed to determine if a model's internal signals are actively used in its decision-making process. Unlike previous methods like the log…
-
New framework enhances power grid stability under uncertainty
Researchers have developed a new framework for ensuring the stability of AC optimal power flow systems under uncertainty. This method leverages a novel geometric approach, transforming stability requirements into convex…
-
New SC-NOs improve neural network accuracy for complex PDE modeling
Researchers have developed Sensitivity-Constrained Neural Operators (SC-NOs) to improve the reliability and data efficiency of neural networks used for modeling partial differential equations (PDEs). By incorporating Ja…
-
New theory explains Jacobian lens for language model interpretation
Researchers have developed a mathematical framework to better understand the Jacobian lens (J-lens), a method used to interpret representations within language models. The study provides a theoretical basis for the J-le…
-
New method enhances data utility under Local Differential Privacy
Researchers have developed a novel method to improve the utility of data collected under Local Differential Privacy (LDP). This approach selectively reduces noise in subspaces of the data that are most relevant to a spe…
-
New research shows normalization layers provide global context in sequence labelers
A new research paper challenges the conventional understanding of context in convolutional sequence labelers. The study demonstrates that sequence-pooled normalization layers can provide global context, bypassing the li…
-
Neural operators fail to reliably warm-start Newton solvers for PDEs
Researchers have identified a critical flaw in using neural operators to warm-start Newton solvers for nonlinear partial differential equations (PDEs). While neural operators can reduce test error, they may still produc…
-
FishBack method improves transformer activation steering using non-Euclidean geometry
Researchers have developed a new method called FishBack to improve activation steering in transformers, a technique for modifying language model behavior without updating parameters. Existing methods are often unstable …
-
New research offers theoretical convergence guarantees for neural network training
Two new research papers explore theoretical underpinnings of neural network training. The first paper establishes convergence guarantees for gradient descent in general feedforward neural networks by introducing a gener…
-
New research proposes unsupervised disentanglement via functional orthogonality
A new research paper proposes a novel approach to unsupervised disentangled representation learning by framing latent concepts as factors influencing observations through locally orthogonal directions. This method, form…
-
Gemma 2-2B research finds active feature planes have less holonomy
A new research paper published on arXiv investigates the concentration of holonomy within specific feature planes of the Gemma 2-2B model. The study preregistered its methodology and analysis rules before inspecting the…
-
Google's Gemma-4-E4B-it model's internal science representations are readable and steerable
Researchers have developed methods to read and steer the internal representations of materials science mechanisms within the open-weight google/gemma-4-E4B-it language model. The study demonstrates that concepts are dis…
-
New research highlights geometry preservation for multimodal contrastive learning
Researchers have identified that the conditioning of encoder Jacobians is crucial for effective trimodal contrastive learning, a method extending beyond simple image-text pairs to align three or more modalities. Poorly …
-
Researchers Explore 'J-space' as Language Model Subconscious
Researchers have explored the concept of a "J-space" within language models, which they liken to the subconscious of these AI systems. This approach, using a "Jacobian lens," allows for a deeper look into the models' in…
-
New Hyperbolic Neural Closure Improves Radiation Transfer Simulations
Researchers have developed a novel hyperbolic neural closure model designed to enhance accuracy and stability in radiation transfer simulations. This new model addresses a critical issue in M1 methods, where unconstrain…
-
Anthropic unveils J-Lens to visualize LLM internal thought processes
Anthropic has introduced a new interpretability technique called the Jacobian Lens (J-Lens) to visualize the internal thought processes of its large language models, specifically Claude. This J-Lens reveals a hidden "J-…
-
Anthropic unveils 'J-space' for Claude AI's internal reasoning
Anthropic has introduced a new internal mechanism for its Claude models called "J-space," which allows the AI to process and store information internally without generating external output. This J-space is described as …
-
New method computes continuous integral R2 indicator using box decomposition
Researchers have developed a new method for computing the continuous integral R2 indicator, a refinement of the classical R2 indicator used in multi-objective optimization and database skyline selection. The approach in…
-
New dual-edge graph enhances interpretable diabetic retinopathy grading
Researchers have developed a novel dual-edge spatial-Jacobian image graph to improve the interpretability of diabetic retinopathy (DR) grading from retinal images. This method represents each fundus photograph as a grap…