Hugging Face Trending Papers

Entangled Representations Amplify Collateral Damage in Unlearning

The paper investigates whether representational entanglement—shared structure between knowledge domains—impedes unlearning in neural networks. By training six 254M‑parameter language models with varying degrees of disentanglement between biology and non‑biology knowledge and applying three unlearning methods, the authors find that more disentangled models consistently achieve better retain‑forget trade‑offs, with up to four‑fold lower retain cost. This controlled experiment provides direct evidence that entanglement contributes to collateral damage during unlearning, supporting a long‑standing hypothesis in interpretability research.

arXiv Machine Learning

Sep 3

Entangled Representations Amplify Collateral Damage in Unlearning

The paper investigates whether representational entanglement—shared structure between knowledge domains—impedes unlearning in neural networks. Using Selective Gradient Masking, the authors train six 254M‑parameter language models with varying degrees of disentanglement between biology and non‑biology knowledge, then apply three standard unlearning methods to each. Results show that more disentangled models consistently achieve better retain‑forget trade‑offs, with up to four‑fold lower retain cost at the same forgetting level, providing direct evidence that entanglement contributes to collateral damage in unlearning.

By Ev\v{z}en Wybitul, Tim G. J. Rudner, Christian Schroeder de Witt

llms safety

arXiv AI

Aug 28

What the "Spotless" Mind Remembers: How Knowledge Entanglement Shapes What Leaks After Unlearning in LLMs

The paper investigates how the structural entanglement of facts within a large language model’s knowledge base influences whether those facts leak after unlearning. Using two unlearning algorithms (WHP and GA+KL) across fictional and real-world datasets and multiple model sizes, the authors find that highly entangled facts are more likely to be recalled before unlearning, but the relationship changes—WHP weakens it while GA+KL reverses it. By directly manipulating entanglement scores and observing corresponding recall changes, they demonstrate a causal link and develop a predictive tool to audit prompts for potential leakage.

By Aakriti Shah, Yifan Hu, Thai Le

llms robotics safety

arXiv Machine Learning

2d ago

Contrastive Representation Shaping for LLM Unlearning

arXiv:2601.22028v2 Announce Type: replace Abstract: Most LLM unlearning methods aim to approximate retrain-from-scratch behaviors with minimal distribution shift, often via alignment-style objectives...

By Haoran Tang, Rajiv Khanna

llms rag benchmarks safety

arXiv Machine Learning

Sep 23

Towards Mitigating Excessive Forgetting in LLM Unlearning via Entanglement-Guidance with Proxy Constraint

arXiv:2508.20443v3 Announce Type: replace Abstract: Large language models (LLMs) are trained on massive datasets that may include private or copyrighted content. Due to growing privacy and ownership...

By Zhihao Liu, Jian Lou, Yuke Hu, Xiaochen Li, Yitian Chen, Tailun Chen, Zhizhen Qin, Kui Ren, Zhan Qin

llms benchmarks safety

arXiv AI

Aug 5

One-Point Contraction: Erasing Representational Separability toward Irreversible Deep Forgetting

arXiv:2507. 07754v3 Announce Type: replace-cross Abstract: Machine unlearning is usually evaluated by what the classifier outputs: forget-set accuracy, confidence, membership-inference scores.

By Jaeheun Jung, Bosung Jung, Suhyun Bae, Donghun Lee

benchmarks

arXiv AI

Jun 6

FIDES: Faithful Inference via Deep Evidence Signals for Retrieval-Memory Conflict in RAG

arXiv:2606. 05644v1 Announce Type: new Abstract: When retrieved evidence contradicts parametric memory, language models frequently ignore context and default to memorized priors -- a failure that undermines the core purpose of retrieval augmentation.

By Zhe Yu, Wenpeng Xing, Tiancheng Zhao, Mohan Li, Changting Lin, Meng Han

llms rag benchmarks safety

arXiv Machine Learning

Jun 15

Natively Unlearnable Large Language Models

arXiv:2606. 13873v1 Announce Type: new Abstract: Unlearning aims to remove the influence of specific training data sources, but this has proved challenging because the contributions of different sources are entangled within the model.

By Gaurav R. Ghosal, Pratyush Maini, Aditi Raghunathan

llms benchmarks safety

arXiv AI

Jul 10

ParamMute: Suppressing Knowledge-Critical FFNs for Faithful Retrieval-Augmented Generation

arXiv:2502. 15543v4 Announce Type: replace-cross Abstract: Large language models (LLMs) integrated with retrieval-augmented generation (RAG) have improved factuality by grounding outputs in external evidence.

By Pengcheng Huang, Zhenghao Liu, Yukun Yan, Haiyan Zhao, Xiaoyuan Yi, Hao Chen, Zhiyuan Liu, Maosong Sun, Tong Xiao, Ge Yu, Chenyan Xiong

llms rag benchmarks

arXiv AI

Aug 11

A Statistical Framework for Auditing Behavioral Dependence and Induced Bias in LLM Judges

arXiv:2604. 07650v2 Announce Type: replace Abstract: The rapid growth of the large language model (LLM) ecosystem raises a critical question: are seemingly diverse models truly independent?

By Chenchen Kuai, Jiwan Jiang, Zihao Zhu, Hao Wang, Keshu Wu, Zihao Li, Yunlong Zhang, Chenxi Liu, Zhengzhong Tu, Zhiwen Fan, Yang Zhou

llms efficiency benchmarks safety

arXiv Machine Learning

4d ago

Do LLMs Really Forget? Hidden-State Leakage in Model Unlearning and How to Fix it

arXiv:2609.36612v1 Announce Type: new Abstract: Unlearning in large language models (LLMs) is typically evaluated at the output level, where a model appears to suppress sensitive or undesirable conte...

By Hadi Reisizadeh, Jiajun Ruan, Sijia Liu, Mingyi Hong

llms benchmarks safety

arXiv AI

Jul 10

Overthinking: Amplifying Reasoning Weights to Extract Learned Secrets

arXiv:2607. 08173v1 Announce Type: new Abstract: Black box auditing of language models is an essential pre-deployment tool, but it may miss subtle forms of misalignment and hidden information.

By Jack Hopkins, Dipika Khullar, Fabien Roger

llms

arXiv AI

Jun 12

From Isolation to Entanglement: When Do Interpretability Methods Identify and Disentangle Known Concepts?

arXiv:2512. 15134v2 Announce Type: replace-cross Abstract: A goal of interpretability is to recover disentangled representations of latent concepts (features) from the activations of neural networks.

By Aaron Mueller, Andrew Lee, Shruti Joshi, Ekdeep Singh Lubana, Dhanya Sridhar, Patrik Reizinger

safety

Related stories