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.
By Pietro Barbiero
arXiv:2604. 07242v3 Announce Type: replace Abstract: Despite deep learning models running well-defined mathematical functions, we lack a formal mathematical framework for describing model architectures.
By Vincent Abbott, Gioele Zardini
arXiv:2608.29034v1 Announce Type: cross
Abstract: A wide range of methods have been proposed for interpreting language models, delivering important insights into their inner workings. However, differ...
By Zhang Enyan, R. Thomas McCoy
arXiv:2608. 01633v1 Announce Type: new Abstract: Large language models (LLMs) enable neural architecture search (NAS) directly over executable neural network programs.
By Zhen Liu, Wanqi Zhou, Shuanghao Bai, Yuhan Liu, Jinjun Wang, Jingwen Fu
Large language models (LLMs) are built from structured high-dimensional objects such as token representations, weights, adaptation updates, caches, and activations, whose multilinear structure is unde...
This survey reviews tensor methods applied to large language models, framing them through a seven‑stage lifecycle (tokenization, embeddings, pre‑training, adaptation, compression, inference, interpretability) and a component view (embeddings, attention, feed‑forward networks). It offers unified notation, theoretical foundations, and comparative analyses of tensorization strategies for Transformer components, while highlighting evaluation protocol differences and model scale effects. The paper also introduces a new metric, ρ_gap, to quantify the gap between theoretical memory savings and actual system‑level speedup, and connects tensor techniques to related efficiency and probabilistic methods.
By Matvei Tarasov, Salman Ahmadi-Asl, Andre L. F. de Almeida, Andrzej Cichocki
The paper discusses tensorizing neural networks by reshaping dense weight matrices into higher-order tensors and approximating them with low-rank tensor network decompositions. This approach offers promising model compression and introduces bond indices that create new latent spaces, potentially enhancing interpretability. Despite encouraging empirical results, tensorized neural networks remain underused, and the authors call for more research to address practical scaling and adoption challenges.
By Safa Hamreras, Sukhbinder Singh, Rom\'an Or\'us
The paper argues that as Large Language Models transition from chatbots to agentic systems, the current post-hoc interpretability paradigm is insufficient for safe deployment because it cannot audit or intervene before an output is produced. It proposes a shift to generative interpretability, where a model’s inference process inherently exposes semantically meaningful checkpoints that are human-understandable and can be causally intervened upon. The authors illustrate the advantages of this approach and introduce Neuro‑Symbolic Models as a concrete implementation.
By Xiaocong Yang
arXiv:2602.11678v2 Announce Type: replace
Abstract: Multimodal Large Language Models (MLLMs) have shown remarkable progress in visual understanding, yet they suffer from a critical limitation: struct...
By Chengwei Ma, Zhen Tian, Zhou Zhou, Zhixian Xu, Xiaowei Zhu, Xia Hua, Si Shi, F. Richard Yu
arXiv:2608. 16773v1 Announce Type: new Abstract: Prototype-based neural networks are hailed as interpretable-by-design architectures.
By Jules Soria, Alban Grastien, Romain Xu-Darme, Julien Girard-Satabin, Zakaria Chihani, Daniela Cancila
arXiv:2609.13198v1 Announce Type: new
Abstract: One of the pivotal recent challenges in neural network interpretability is polysemanticity, where a single neuron is activated by multiple, often unrel...
By Sehyun Lee, Dahee Kwon, Damin Lee, Jaesik Choi
Neuton is a framework that automatically converts neuron‑level abstract interpreter specifications into tensor‑level executables, enabling the creation of new DNN certifiers without manual tensor programming. It introduces a stack‑based intermediate representation with shape analysis, domain‑specific rewrite rules for efficient tensor operations, and a novel sparse tensor format (gBCSR) tailored to certifier sparsity. The authors demonstrate that Neuton preserves neuron‑level semantics and evaluate it on eight certifiers across 32 DNN benchmarks, compiling four existing certifiers and producing executable implementations for four others, revealing new precision‑runtime trade‑offs such as SkipPoly matching DeepPoly’s precision while often running faster.
By Avaljot Singh, Yasmin Chandini Sarita, Aditya Mishra, Gagandeep Singh, Charith Mendis