← Back to all news
arXiv Computation and Language By Ahmad Dawar Hakimi, Kaiwei Lei, Isabelle Augenstein, Hinrich Sch\"utze

Output Vector Editing for Memorization Mitigation in Large Language Models

Read the original on arXiv Computation and Language →

Output Vector Editing (OVE) is a post‑hoc technique that suppresses unwanted text continuations from large language models by editing the output vectors of specific neurons. It uses a two‑stage localization of MLP layers and neurons combined with full‑vocabulary least‑squares updates, and its multi‑anchor variant (OVE‑MA) adaptively selects continuation positions to maximize suppression. Experiments on models ranging from 360 M to 7 B parameters show that OVE‑MA can reduce recoverable continuations by 24.45 % with only five output‑vector updates, outperforming the prior method REVS while requiring far fewer updates.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computation and Language.

Related stories

arXiv Machine Learning
Oct 16

Mitigating Memorization Where It Happens

The paper introduces a token‑level log‑likelihood ratio to pinpoint where large language models memorize training sequences. It proposes a probe‑steer framework, instantiated as Gated Subspace Steering (GSS), that selectively corrects memorization‑relevant activations using the leading singular vectors of a gradient‑weighted activation matrix. Experiments on four benchmarks show GSS matches or surpasses state‑of‑the‑art memorization reduction methods while using 10–100× less compute.

By Xuanqi Zhang, Haoyang Shang, Xiaoxiao Li
llmsbenchmarkssafety
More like this →
arXiv AI
Oct 1

Mitigating Memorization In Language Models

The paper explores ways to reduce the memorization of training data in language models, testing three regularizer-based, three finetuning-based, and eleven machine unlearning methods—five of which are newly introduced. It introduces TinyMem, a lightweight suite of small models for rapid testing of these mitigation techniques, and shows that unlearning methods, particularly BalancedSubnet, outperform others in removing memorized content while maintaining task performance. The study also finds that regularizer-based approaches are slow and ineffective, while finetuning methods are costly, especially when high accuracy is required.

By Mansi Sakarvadia, Aswathy Ajith, Arham Khan, Nathaniel Hudson, Caleb Geniesse, Kyle Chard, Yaoqing Yang, Ian Foster, Michael W. Mahoney
llmsfine-tuning
More like this →
arXiv AI
Sep 25

The Tokens Remember: When Tokenization Bypasses Knowledge Editing and Unlearning

The paper investigates how tokenization can undermine post‑release guarantees that sensitive knowledge has been edited or unlearned from open‑weight large language models. By showing that alternative valid tokenizations can bypass localized modifications, the authors introduce Toketive, a reference‑free attack that detects modified knowledge and reconstructs pre‑edit responses using only the released model. Experiments on five LLMs, six datasets, and six editing techniques reveal that 38.6% of alternative tokenizations recover suppressed information, with Toketive achieving high detection and reconstruction accuracy.

By Manit Baser, Aditya Nawal, Dinil Mon Divakaran, Mohan Gurusamy
llmsnlpsafety
More like this →