Copertina

Why Deep Jacobian Spectra Separate: Depth-Induced Scaling and Singular-Vector Alignment

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Haas, Nathanaël, Gatine, François, Cosse, Augustin M, Bouraoui, Zied
Natura:	Preprint
Pubblicazione:	2026
Soggetti:	Machine Learning Artificial Intelligence
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

Credal Concept Bottleneck Models: Structural Separation of Epistemic and Aleatoric Uncertainty
di: Mukherjee, Tanmoy, et al.
Pubblicazione: (2026)

Vector Field Oriented Diffusion Model for Crystal Material Generation
di: Klipfel, Astrid, et al.
Pubblicazione: (2023)

MODE: Multi-Objective Adaptive Coreset Selection
di: Mukherjee, Tanmoy, et al.
Pubblicazione: (2025)

Generalizing Analogical Inference from Boolean to Continuous Domains
di: Cunha, Francisco, et al.
Pubblicazione: (2025)

Fourier Transformers for Latent Crystallographic Diffusion and Generative Modeling
di: Duersch, Jed A., et al.
Pubblicazione: (2026)

Grounding Agent Reasoning in Image Schemas: A Neurosymbolic Approach to Embodied Cognition
di: Olivier, François, et al.
Pubblicazione: (2025)

Towards a Neurosymbolic Reasoning System Grounded in Schematic Representations
di: Olivier, François, et al.
Pubblicazione: (2025)

Singular Vectors of Attention Heads Align with Features
di: Franco, Gabriel, et al.
Pubblicazione: (2026)

PiSSA: Principal Singular Values and Singular Vectors Adaptation of Large Language Models
di: Meng, Fanxu, et al.
Pubblicazione: (2024)

Any-Depth Alignment: Unlocking Innate Safety Alignment of LLMs to Any-Depth
di: Zhang, Jiawei, et al.
Pubblicazione: (2025)

VectorFit : Adaptive Singular & Bias Vector Fine-Tuning of Pre-trained Foundation Models
di: Hegde, Suhas G, et al.
Pubblicazione: (2025)

Multimodal Deep Learning for Low-Resource Settings: A Vector Embedding Alignment Approach for Healthcare Applications
di: Restrepo, David, et al.
Pubblicazione: (2024)

The Depth Delusion: Why Transformers Should Be Wider, Not Deeper
di: Fahim, Md Muhtasim Munif, et al.
Pubblicazione: (2026)

Skim-Aware Contrastive Learning for Efficient Document Representation
di: Abro, Waheed Ahmed, et al.
Pubblicazione: (2025)

Adaptive Helpfulness-Harmlessness Alignment with Preference Vectors
di: Liang, Ren-Wei, et al.
Pubblicazione: (2025)

Deep Support Vectors
di: Lee, Junhoo, et al.
Pubblicazione: (2024)

SVFT: Parameter-Efficient Fine-Tuning with Singular Vectors
di: Lingam, Vijay, et al.
Pubblicazione: (2024)

Transformers Linearly Represent Highly Structured World Models
di: Kniazev, Roman, et al.
Pubblicazione: (2026)

Beyond Components: Singular Vector-Based Interpretability of Transformer Circuits
di: Ahmad, Areeb, et al.
Pubblicazione: (2025)

Dynamic Latent Separation for Deep Learning
di: Tuan, Yi-Lin, et al.
Pubblicazione: (2022)

Spectra 1.1: Scaling Laws and Efficient Inference for Ternary Language Models
di: Vaidhya, Tejas, et al.
Pubblicazione: (2025)

MIDUS: Memory-Infused Depth Up-Scaling
di: Kim, Taero, et al.
Pubblicazione: (2025)

Position: Capability Control Should be a Separate Goal From Alignment
di: Siddiqui, Shoaib Ahmed, et al.
Pubblicazione: (2026)

Not Just RLHF: Why Alignment Alone Won't Fix Multi-Agent Sycophancy
di: Kumarappan, Adarsh, et al.
Pubblicazione: (2026)

LLM-VA: Resolving the Jailbreak-Overrefusal Trade-off via Vector Alignment
di: Zhang, Haonan, et al.
Pubblicazione: (2026)

Spectra-to-Structure and Structure-to-Spectra Inference Across the Periodic Table
di: Wang, Yufeng, et al.
Pubblicazione: (2025)

Parameter-Efficient Neural CDEs via Implicit Function Jacobians
di: Kuleshov, Ilya, et al.
Pubblicazione: (2025)

Mitigating Adversarial Perturbations for Deep Reinforcement Learning via Vector Quantization
di: Luu, Tung M., et al.
Pubblicazione: (2024)

Deep-Learning Investigation of Vibrational Raman Spectra for Plant-Stress Analysis
di: Patil, Anoop C., et al.
Pubblicazione: (2025)

Causality--Δ: Jacobian-Based Dependency Analysis in Flow Matching Models
di: Rezvan, Reza, et al.
Pubblicazione: (2026)

ANCRe: Adaptive Neural Connection Reassignment for Efficient Depth Scaling
di: Zhang, Yilang, et al.
Pubblicazione: (2026)

Jacobian Descent for Multi-Objective Optimization
di: Quinton, Pierre, et al.
Pubblicazione: (2024)

Fractional-order Jacobian Matrix Differentiation and Its Application in Artificial Neural Networks
di: zhou, Xiaojun, et al.
Pubblicazione: (2025)

Negligible in Size, Significant in Effect: On Scale Vectors in Large Language Models
di: Wang, Mingze, et al.
Pubblicazione: (2026)

Credal Concept Bottleneck Models for Epistemic-Aleatoric Uncertainty Decomposition
di: Mukherjee, Tanmoy, et al.
Pubblicazione: (2026)

Representational Alignment with Chemical Induced Fit for Molecular Relational Learning
di: Zhang, Peiliang, et al.
Pubblicazione: (2025)

Generic Multi-modal Representation Learning for Network Traffic Analysis
di: Gioacchini, Luca, et al.
Pubblicazione: (2024)

Episodic Memories Generation and Evaluation Benchmark for Large Language Models
di: Huet, Alexis, et al.
Pubblicazione: (2025)

Position: AI Evaluations Should be Grounded on a Theory of Capability
di: Jo, Nathanael, et al.
Pubblicazione: (2025)

The Alignment Problem from a Deep Learning Perspective
di: Ngo, Richard, et al.
Pubblicazione: (2022)