GPT-2 Through the Lens of Vector Symbolic Architectures
Understanding the general priniciples behind transformer models remains a complex endeavor. Experiments with probing and disentangling features using sparse autoencoders (SAE) suggest that these models might manage linear features embedded as directions in the residual stream. This paper explores the resemblance between decoder-only transformer architecture and vector symbolic architectures (VSA) and presents experiments indicating that GPT-2 uses mechanisms involving nearly orthogonal vector bundling and binding operations similar to VSA for computation and communication between layers. It further shows that these principles help explain a significant portion of the actual neural weights.
Code (0)
등록된 구현이 없습니다.
Tasks
DecoderMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
Bridging Compositional and Distributional Semantics: A Survey on Latent Semantic Geometry via AutoEncoder
Integrating compositional and symbolic properties into current distributional semantic spaces can enhance the interpretability, controllability, compositionality, and generalisation capabilities of Transformer-based auto…
Representation LearningProbabilistic Abduction for Visual Abstract Reasoning via Learning Rules in Vector-symbolic Architectures
Abstract reasoning is a cornerstone of human intelligence, and replicating it with artificial intelligence (AI) presents an ongoing challenge. This study focuses on efficiently solving Raven's progressive matrices (RPM),…
AttributeBridge Networks: Relating Inputs through Vector-Symbolic Manipulations
Despite rapid progress, current deep learning methods face a number of critical challenges. These include high energy consumption, catastrophic forgetting, dependance on global losses, and an inability to reason symbolic…
SymFlux: deep symbolic regression of Hamiltonian vector fields
We present SymFlux, a novel deep learning framework that performs symbolic regression to identify Hamiltonian functions from their corresponding vector fields on the standard symplectic plane. SymFlux models utilize hybr…
A Walsh Hadamard Derived Linear Vector Symbolic Architecture
Vector Symbolic Architectures (VSAs) are one approach to developing Neuro-symbolic AI, where two vectors in $\mathbb{R}^d$ are `bound' together to produce a new vector in the same space. VSAs support the commutativity an…
Computational Efficiency