Copertina

GSS: Gated Subspace Steering for Selective Memorization Mitigation in LLMs

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Zhang, Xuanqi, Shang, Haoyang, Li, Xiaoxiao
Natura:	Preprint
Pubblicazione:	2026
Soggetti:	Machine Learning
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

The Landscape of Memorization in LLMs: Mechanisms, Measurement, and Mitigation
di: Xiong, Alexander, et al.
Pubblicazione: (2025)

Subspace Control: Turning Constrained Model Steering into Controllable Spectral Optimization
di: Huang, Yancheng, et al.
Pubblicazione: (2026)

Understanding and Mitigating Memorization in Diffusion Models for Tabular Data
di: Fang, Zhengyu, et al.
Pubblicazione: (2024)

Mitigating Unintended Memorization with LoRA in Federated Learning for LLMs
di: Bossy, Thierry, et al.
Pubblicazione: (2025)

How Do Flow Matching Models Memorize and Generalize in Sample Data Subspaces?
di: Gao, Weiguo, et al.
Pubblicazione: (2024)

Memorizing Long-tail Data Can Help Generalization Through Composition
di: Zhou, Mo, et al.
Pubblicazione: (2025)

Randomized Masked Finetuning: An Efficient Way to Mitigate Memorization of PIIs in LLMs
di: Joshi, Kunj, et al.
Pubblicazione: (2025)

Localizing and Mitigating Memorization in Image Autoregressive Models
di: Kasliwal, Aditya, et al.
Pubblicazione: (2025)

Mitigating Memorization In Language Models
di: Sakarvadia, Mansi, et al.
Pubblicazione: (2024)

SteerConf: Steering LLMs for Confidence Elicitation
di: Zhou, Ziang, et al.
Pubblicazione: (2025)

Interpreting and Steering State-Space Models via Activation Subspace Bottlenecks
di: Mohan, Vamshi Sunku, et al.
Pubblicazione: (2026)

Shifting Perspectives: Steering Vectors for Robust Bias Mitigation in LLMs
di: Siddique, Zara, et al.
Pubblicazione: (2025)

Rethinking LoRA for Data Heterogeneous Federated Learning: Subspace and State Alignment
di: Peng, Hongyi, et al.
Pubblicazione: (2026)

Memorized Images in Diffusion Models share a Subspace that can be Located and Deleted
di: Chavhan, Ruchika, et al.
Pubblicazione: (2024)

SafeSteer: Interpretable Safety Steering with Refusal-Evasion in LLMs
di: Ghosh, Shaona, et al.
Pubblicazione: (2025)

Memorization vs. Reasoning: Updating LLMs with New Knowledge
di: Li, Aochong Oliver, et al.
Pubblicazione: (2025)

TARDIS: Mitigating Temporal Misalignment via Representation Steering
di: Shin, Changho, et al.
Pubblicazione: (2025)

Understanding and Mitigating Memorization in Generative Models via Sharpness of Probability Landscapes
di: Jeon, Dongjae, et al.
Pubblicazione: (2024)

Where to Steer: Input-Dependent Layer Selection for Steering Improves LLM Alignment
di: Gadgil, Soham, et al.
Pubblicazione: (2026)

Mitigating Memorization in LLMs using Activation Steering
di: Suri, Manan, et al.
Pubblicazione: (2025)

Conjunction Subspaces Test for Conformal and Selective Classification
di: He, Zengyou, et al.
Pubblicazione: (2024)

Steering LLMs via Scalable Interactive Oversight
di: Zhou, Enyu, et al.
Pubblicazione: (2026)

FineGates: LLMs Finetuning with Compression using Stochastic Gates
di: Svirsky, Jonathan, et al.
Pubblicazione: (2024)

(Token-Level) InfoRMIA: Stronger Membership Inference and Memorization Assessment for LLMs
di: Tao, Jiashu, et al.
Pubblicazione: (2025)

Zeroth-Order Fine-Tuning of LLMs in Random Subspaces
di: Yu, Ziming, et al.
Pubblicazione: (2024)

Spurious Rewards Paradox: Mechanistically Understanding How RLVR Activates Memorization Shortcuts in LLMs
di: Yan, Lecheng, et al.
Pubblicazione: (2026)

Captured by Captions: On Memorization and its Mitigation in CLIP Models
di: Wang, Wenhao, et al.
Pubblicazione: (2025)

In-Context Molecular Property Prediction with LLMs: A Blinding Study on Memorization and Knowledge Conflicts
di: Busch, Matthias, et al.
Pubblicazione: (2026)

Memorize Early, Then Query: Inlier-Memorization-Guided Active Outlier Detection
di: Kang, Minseo, et al.
Pubblicazione: (2026)

Steering Away from Memorization: Reachability-Constrained Reinforcement Learning for Text-to-Image Diffusion
di: Karnik, Sathwik, et al.
Pubblicazione: (2026)

The Pitfalls of Memorization: When Memorization Hurts Generalization
di: Bayat, Reza, et al.
Pubblicazione: (2024)

MemHunter: Automated and Verifiable Memorization Detection at Dataset-scale in LLMs
di: Wu, Zhenpeng, et al.
Pubblicazione: (2024)

Gated Subspace Inference for Transformer Acceleration
di: Thomas, Stephen J.
Pubblicazione: (2026)

Sub-MoE: Efficient Mixture-of-Expert LLMs Compression via Subspace Expert Merging
di: Li, Lujun, et al.
Pubblicazione: (2025)

Selective Safety Steering via Value-Filtered Decoding
di: Einbinder, Bat-Sheva, et al.
Pubblicazione: (2026)

Token Hidden Reward: Steering Exploration-Exploitation in Group Relative Deep Reinforcement Learning
di: Deng, Wenlong, et al.
Pubblicazione: (2025)

Mitigating Cognitive Inertia in Large Reasoning Models via Latent Spike Steering
di: Lee, Seojin, et al.
Pubblicazione: (2026)

Understanding and Mitigating Dataset Corruption in LLM Steering
di: Anderson, Cullen, et al.
Pubblicazione: (2026)

Exploring the Linear Subspace Hypothesis in Gender Bias Mitigation
di: Vargas, Francisco, et al.
Pubblicazione: (2020)

DSO: Direct Steering Optimization for Bias Mitigation
di: Paes, Lucas Monteiro, et al.
Pubblicazione: (2025)