Copertina

Superposition Is Not Necessary: A Mechanistic Interpretability Analysis of Transformer Representations for Time Series Forecasting

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autore principale:	Yıldırım, Alper
Natura:	Preprint
Pubblicazione:	2026
Soggetti:	Machine Learning Artificial Intelligence I.2.6
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

The Geometric Inductive Bias of Grokking: Bypassing Phase Transitions via Architectural Topology
di: Yıldırım, Alper
Pubblicazione: (2026)

Language as a Wave Phenomenon: Semantic Phase Locking and Interference in Neural Networks
di: Yıldırım, Alper, et al.
Pubblicazione: (2025)

DRFormer: Multi-Scale Transformer Utilizing Diverse Receptive Fields for Long Time-Series Forecasting
di: Ding, Ruixin, et al.
Pubblicazione: (2024)

Evaluating the effectiveness of predicting covariates in LSTM Networks for Time Series Forecasting
di: Davies, Gareth
Pubblicazione: (2024)

Inter-Series Transformer: Attending to Products in Time Series Forecasting
di: Cristian, Rares, et al.
Pubblicazione: (2024)

Robust Multivariate Time Series Forecasting against Intra- and Inter-Series Transitional Shift
di: He, Hui, et al.
Pubblicazione: (2024)

FreRA: A Frequency-Refined Augmentation for Contrastive Learning on Time Series Classification
di: Tian, Tian, et al.
Pubblicazione: (2025)

Deep Memory Search: A Metaheuristic Approach for Optimizing Heuristic Search
di: Hedar, Abdel-Rahman, et al.
Pubblicazione: (2024)

TS-ACL: Closed-Form Solution for Time Series-oriented Continual Learning
di: Li, Jiaxu, et al.
Pubblicazione: (2024)

Black Box Model Explanations and the Human Interpretability Expectations -- An Analysis in the Context of Homicide Prediction
di: Ribeiro, José, et al.
Pubblicazione: (2022)

Noise or Signal? Deconstructing Contradictions and An Adaptive Remedy for Reversible Normalization in Time Series Forecasting
di: Fu, Fanzhe, et al.
Pubblicazione: (2025)

DecompKAN: Decomposed Patch-KAN for Long-Term Time Series Forecasting
di: Mysore, Naveen
Pubblicazione: (2026)

Load and Renewable Energy Forecasting Using Deep Learning for Grid Stability
di: Sarkar, Kamal
Pubblicazione: (2025)

Entropy Causal Graphs for Multivariate Time Series Anomaly Detection
di: Febrinanto, Falih Gozi, et al.
Pubblicazione: (2023)

A Set-Sequence Model for Time Series
di: Epstein, Elliot L., et al.
Pubblicazione: (2025)

GCAD: Anomaly Detection in Multivariate Time Series from the Perspective of Granger Causality
di: Liu, Zehao, et al.
Pubblicazione: (2025)

A Learnable Multi-views Contrastive Framework with Reconstruction Discrepancy for Medical Time-Series
di: Wang, Yifan, et al.
Pubblicazione: (2025)

The Deterministic Horizon: When Extended Reasoning Fails and Tool Delegation Becomes Necessary
di: Guo, Dongxin, et al.
Pubblicazione: (2026)

Interpretability-Guided Bi-objective Optimization: Aligning Accuracy and Explainability
di: Fouladi, Kasra, et al.
Pubblicazione: (2026)

A Hybrid Model for Stock Market Forecasting: Integrating News Sentiment and Time Series Data with Graph Neural Networks
di: Sadek, Nader, et al.
Pubblicazione: (2025)

Prototype Transformer: Towards Language Model Architectures Interpretable by Design
di: Yordanov, Yordan, et al.
Pubblicazione: (2026)

Causal Discovery in Semi-Stationary Time Series
di: Gao, Shanyun, et al.
Pubblicazione: (2024)

Generative and Contrastive Graph Representation Learning
di: Chen, Jiali, et al.
Pubblicazione: (2025)

Decentralized Time Series Classification with ROCKET Features
di: Casella, Bruno, et al.
Pubblicazione: (2025)

Position: Mechanistic Interpretability Must Disclose Identification Assumptions for Causal Claims
di: Lin, Zezheng, et al.
Pubblicazione: (2026)

One Router to Route Them All: Homogeneous Expert Routing for Heterogeneous Graph Transformers
di: Shakirov, Georgiy, et al.
Pubblicazione: (2025)

Weakly Supervised Distillation of Hallucination Signals into Transformer Representations
di: Salehmohamed, Shoaib Sadiq, et al.
Pubblicazione: (2026)

LaT-PFN: A Joint Embedding Predictive Architecture for In-context Time-series Forecasting
di: Verdenius, Stijn, et al.
Pubblicazione: (2024)

Low-Dimensional Execution Manifolds in Transformer Learning Dynamics: Evidence from Modular Arithmetic Tasks
di: Xu, Yongzhong
Pubblicazione: (2026)

Causal Discovery-Driven Change Point Detection in Time Series
di: Gao, Shanyun, et al.
Pubblicazione: (2024)

TimeCatcher: A Variational Framework for Volatility-Aware Forecasting of Non-Stationary Time Series
di: Chen, Zhiyu, et al.
Pubblicazione: (2026)

Time-to-Injury Forecasting in Elite Female Football: A DeepHit Survival Approach
di: Catterall, Victoria, et al.
Pubblicazione: (2026)

Shattered Compositionality: Counterintuitive Learning Dynamics of Transformers for Arithmetic
di: Zhao, Xingyu, et al.
Pubblicazione: (2026)

On the Role of Pre-trained Embeddings in Binary Code Analysis
di: Maier, Alwin, et al.
Pubblicazione: (2025)

A Self-explainable Model of Long Time Series by Extracting Informative Structured Causal Patterns
di: Wang, Ziqian, et al.
Pubblicazione: (2025)

RHiOTS: A Framework for Evaluating Hierarchical Time Series Forecasting Algorithms
di: Roque, Luis, et al.
Pubblicazione: (2024)

Task-Conditioned Routing Signatures in Sparse Mixture-of-Experts Transformers
di: Avinash, Mynampati Sri Ranganadha
Pubblicazione: (2026)

NoRIN: Backbone-Adaptive Reversible Normalization for Time-Series Forecasting
di: Zhang, Shun, et al.
Pubblicazione: (2026)

What Do World Models Learn in RL? Probing Latent Representations in Learned Environment Simulators
di: Zhang, Xinyu
Pubblicazione: (2026)

An Idiosyncrasy of Time-discretization in Reinforcement Learning
di: De Asis, Kris, et al.
Pubblicazione: (2024)