Imagen de Portada

Rethinking LLM Unlearning Objectives: A Gradient Perspective and Go Beyond

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Wang, Qizhou, Zhou, Jin Peng, Zhou, Zhanke, Shin, Saebyeol, Han, Bo, Weinberger, Kilian Q.
Formato:	Preprint
Publicado:	2025
Materias:	Machine Learning
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

LLM Unlearning with LLM Beliefs
por: Li, Kemou, et al.
Publicado: (2025)

Music Transcription with (Almost) No Supervision
por: Shin, Saebyeol, et al.
Publicado: (2026)

Is Gradient Ascent Really Necessary? Memorize to Forget for Machine Unlearning
por: Huang, Zhuo, et al.
Publicado: (2026)

Towards Effective Evaluations and Comparisons for LLM Unlearning Methods
por: Wang, Qizhou, et al.
Publicado: (2024)

Exploring Criteria of Loss Reweighting to Enhance LLM Unlearning
por: Yang, Puning, et al.
Publicado: (2025)

Don't Trust: Verify -- Grounding LLM Quantitative Reasoning with Autoformalization
por: Zhou, Jin Peng, et al.
Publicado: (2024)

Go Beyond Your Means: Unlearning with Per-Sample Gradient Orthogonalization
por: Shamsian, Aviv, et al.
Publicado: (2025)

$Q\sharp$: Provably Optimal Distributional RL for LLM Post-Training
por: Zhou, Jin Peng, et al.
Publicado: (2025)

From Debate to Equilibrium: Belief-Driven Multi-Agent LLM Reasoning via Bayesian Nash Equilibrium
por: Yi, Xie, et al.
Publicado: (2025)

Langevin Unlearning: A New Perspective of Noisy Gradient Descent for Machine Unlearning
por: Chien, Eli, et al.
Publicado: (2024)

Model Inversion Attacks: A Survey of Approaches and Countermeasures
por: Zhou, Zhanke, et al.
Publicado: (2024)

Explainable LLM Unlearning Through Reasoning
por: Liao, Junfeng, et al.
Publicado: (2026)

Distinguishable Deletion: Unifying Knowledge Erasure and Refusal for Large Language Model Unlearning
por: Yang, Puning, et al.
Publicado: (2026)

Reference-guided Policy Optimization for Molecular Optimization via LLM Reasoning
por: Li, Xuan, et al.
Publicado: (2026)

GRU: Mitigating the Trade-off between Unlearning and Retention for LLMs
por: Wang, Yue, et al.
Publicado: (2025)

Graders should cheat: privileged information enables expert-level automated evaluations
por: Zhou, Jin Peng, et al.
Publicado: (2025)

Understanding Fairness Surrogate Functions in Algorithmic Fairness
por: Yao, Wei, et al.
Publicado: (2023)

Envisioning Outlier Exposure by Large Language Models for Out-of-Distribution Detection
por: Cao, Chentao, et al.
Publicado: (2024)

Graph Federated Unlearning for Privacy Preservation
por: Ma, Ruotong, et al.
Publicado: (2026)

Can Language Models Perform Robust Reasoning in Chain-of-thought Prompting with Noisy Rationales?
por: Zhou, Zhanke, et al.
Publicado: (2024)

Efficient Utility-Preserving Machine Unlearning with Implicit Gradient Surgery
por: Zhou, Shiji, et al.
Publicado: (2025)

Less is More: One-shot Subgraph Reasoning on Large-scale Knowledge Graphs
por: Zhou, Zhanke, et al.
Publicado: (2024)

Diffusion Guided Language Modeling
por: Lovelace, Justin, et al.
Publicado: (2024)

DeepInception: Hypnotize Large Language Model to Be Jailbreaker
por: Li, Xuan, et al.
Publicado: (2023)

Neural Atoms: Propagating Long-range Interaction in Molecular Graphs through Efficient Communication Channel
por: Li, Xuan, et al.
Publicado: (2023)

Beyond Single-Value Metrics: Evaluating and Enhancing LLM Unlearning with Cognitive Diagnosis
por: Lang, Yicheng, et al.
Publicado: (2025)

Noisy Test-Time Adaptation in Vision-Language Models
por: Cao, Chentao, et al.
Publicado: (2025)

Gradient-Based Multi-Objective Deep Learning: Algorithms, Theories, Applications, and Beyond
por: Chen, Weiyu, et al.
Publicado: (2025)

Towards LLM Unlearning Resilient to Relearning Attacks: A Sharpness-Aware Minimization Perspective and Beyond
por: Fan, Chongyu, et al.
Publicado: (2025)

Rethinking LLM Ensembling from the Perspective of Mixture Models
por: Fu, Jiale, et al.
Publicado: (2026)

AEGIS: Adversarial Target-Guided Retention-Data-Free Robust Concept Erasure from Diffusion Models
por: Li, Fengpeng, et al.
Publicado: (2026)

From Passive to Active Reasoning: Can Large Language Models Ask the Right Questions under Incomplete Information?
por: Zhou, Zhanke, et al.
Publicado: (2025)

Mind the Gap Between Prototypes and Images in Cross-domain Finetuning
por: Tian, Hongduan, et al.
Publicado: (2024)

SpeechOp: Inference-Time Task Composition for Generative Speech Processing
por: Lovelace, Justin, et al.
Publicado: (2025)

Targeted Unlearning with Single Layer Unlearning Gradient
por: Cai, Zikui, et al.
Publicado: (2024)

Prescriptive Scaling Laws for Data Constrained Training
por: Lovelace, Justin, et al.
Publicado: (2026)

Online Feature Updates Improve Online (Generalized) Label Shift Adaptation
por: Wu, Ruihan, et al.
Publicado: (2024)

Rethinking Multi-Modal Learning from Gradient Uncertainty
por: Guo, Peizheng, et al.
Publicado: (2025)

Label Smoothing Improves Gradient Ascent in LLM Unlearning
por: Pang, Zirui, et al.
Publicado: (2025)

Adaptive Moments are Surprisingly Effective for Plug-and-Play Diffusion Sampling
por: Belardi, Christian, et al.
Publicado: (2026)