Large Language Models (LLMs) drive current AI breakthroughs despite very little being known about their internal representations. In this work, we propose to shed the light on LLMs inner mechanisms through the lens of geometry. In particular, we develop in closed form $(i)$ the intrinsic dimension in which the Multi-Head Attention embeddings are constrained to exist and $(ii)$ the partition and per-region affine mappings of the feedforward (MLP) network of LLMs’ layers. Our theoretical findings further enable the design of novel principled solutions applicable to state-of-the-art LLMs. First, we show that, through our geometric understanding, we can bypass LLMs’ RLHF protection by controlling the embedding’s intrinsic dimension through informed prompt manipulation. Second, we derive interpretable geometrical features that can be extracted from any (pre-trained) LLM, providing a rich abstract representation of their inputs. We observe that these features are sufficient to help solve toxicity detection, and even allow the identification of various types of toxicity. Our results demonstrate how, even in large-scale regimes, exact theoretical results can answer practical questions in LLMs. Code: https://github.com/RandallBalestriero/SplineLLM

We propose a novel approach to filter bank learning for time-series by considering spectral decompositions of signals defined as a Group Transform. This framework allows us to generalize classical time-frequency transformations such as the Wavelet Transform, and to efficiently learn the representation of signals. While the creation of the wavelet transform filter-bank relies on affine transformations of a mother filter, our approach allows for non-linear transformations. The transformations induced by such maps enable us to span a larger class of signal representations, from wavelet to chirplet-like filters. We propose a parameterization of such a non-linear map such that its sampling can be optimized for a specific task and signal. The Learnable Group Transform can be cast into a Deep Neural Network. The experiments on diverse time-series datasets demonstrate the expressivity of this framework, which competes with state-of-the-art performances.

We provide the first frame agnostic thresholding scheme based on risk minimization, which can be applied to arbitrary frames and provide its theoretical guarantees. We investigate the proposed scheme, study its empirical risk, and demonstrates how it falls back to the standard Donoho thresholding scheme for orthogonal basis. We then validate our technique and apply it to the overcomplete wavelet transforms of the Deep Scattering Network. We are thus obtaining an invariant and thresholded representation of the signals providing significant performance gains compared to the non-thresholded version.

We study the geometry of deep (neural) networks (DNs) with piecewise affine and convex nonlinearities. The layers of such DNs have been shown to be max-affine spline operators (MASOs) that partition their input space and apply a region-dependent affine mapping to their input to produce their output. We demonstrate that each MASO layer's input space partitioning corresponds to a power diagram (an extension of the classical Voronoi tiling) with a number of regions that grows exponentially with respect to the number of units (neurons). We further show that a composition of MASO layers (e.g., the entire DN) produces a progressively subdivided power diagram and provide its analytical form. The subdivision process constrains the affine maps on the potentially exponentially many power diagram regions with respect to the number of neurons to greatly reduce their complexity. For classification problems, we obtain a formula for a MASO DN's decision boundary in the input space plus a measure of its curvature that depends on the DN's nonlinearities, weights, and architecture. Numerous numerical experiments support and extend our theoretical results.

We propose to tackle the problem of end-to-end learning for raw waveform signals by introducing learnable continuous time-frequency atoms. The derivation of these filters is achieved by defining a functional space with a given smoothness order and boundary conditions. From this space, we derive the parametric analytical filters. Their differentiability property allows gradient-based optimization. As such, one can utilize any Deep Neural Network (DNN) with these filters. This enables us to tackle in a front-end fashion a large scale bird detection task based on the freefield1010 dataset known to contain key challenges, such as the dimensionality of the inputs data ($>100,000$) and the presence of additional noises: multiple sources and soundscapes.