arXiv:2509. 16525v2 Announce Type: replace-cross Abstract: Machine learning models are increasingly deployed as software components that must evolve as requirements change.
arXiv:2611.06681v1 Announce Type: new Abstract: Machine unlearning is commonly evaluated with standard post-deletion metrics that compare an unlearned model to one retrained on the remaining data. Su...
arXiv:2604. 26866v2 Announce Type: replace-cross Abstract: Large language models (LLMs) acquire most of their factual knowledge during the pre-training stage, through next token prediction.
The paper introduces Update‑Matched Unlearning (UMU), a method that applies the same simulated relearning update to both a defended model and an undefended reference to better preserve plasticity while resisting relearning of deleted knowledge. Experiments on Llama and Qwen models with TOFU and FaithUn datasets show that UMU reduces harmful and benign relearning by 30.4 and 9.9 points on average and increases new‑task accuracy by 3.5 points, supporting the reference‑timing hypothesis. The study demonstrates that robust unlearning should selectively prevent recovery of deleted knowledge without suppressing the model’s ability to learn new tasks.
The paper investigates how machine unlearning for large language models (LLMs) can unintentionally erase related knowledge, even in distant domains. By analyzing the propagation of unlearning effects before any model updates, the authors discover a consistent decay pattern where collateral damage is strongest near the targeted forget set and diminishes with semantic distance but never fully disappears at domain boundaries. They propose a pre-unlearning prediction task—forget-set auditing—to identify potential collateral damage early, finding that interaction features between the forget set and evaluation set are the most predictive signals. This approach offers an early warning system for risky unlearning runs and guides the design of more reliable unlearning procedures.
arXiv:2607. 25546v1 Announce Type: new Abstract: Given a model that is already trained, which features does it rely on causally versus spuriously?
The paper audits whether synthetic distractors in RLVR corpora act as shortcuts for learning policies. A classifier using only surface statistics barely outperforms chance, and manual inspection reveals that code distractors are almost identical to correct answers. Experiments with a paraphrase‑matched control show no exploitation advantage for the unmodified data, indicating that the detectable artifact was not used by the policy.
The paper investigates subject-level machine unlearning as a post‑hoc sanitization technique for engagement recognition datasets that are subject‑indexed and often noisy. By ranking harmful subjects with a model‑dependent proxy and applying a lightweight approximate unlearning update, the authors compare the resulting model to an oracle retrained from scratch on the remaining subjects. Experiments on DAiSEE and EngageNet show that the unlearned model recovers roughly 90–93% of the oracle’s performance at about one‑quarter the retraining cost, with effectiveness peaking at intermediate forget‑set sizes.
arXiv:2606. 16110v1 Announce Type: new Abstract: Machine unlearning has been extensively studied in response to growing privacy concerns and regulatory requirements.
arXiv:2607. 19442v1 Announce Type: cross Abstract: Machine unlearning is commonly evaluated by matching a retrained oracle on trained probes.
The paper presents a certified rule that determines whether unseen data should be incorporated into a pretrained model’s context or its weights, based on an exact law for their information value. This law quantifies the code length saved on held‑out queries as the expected bits stored about the query, adjusted by a prior repair and decoding loss, and applies to any model, architecture, dataset, or task. The authors demonstrate that the rule’s ceilings correctly certify choices across diverse models—including transformers, TabPFN, occupancy networks, and stereo networks—and that an anytime‑valid test can reliably select the better approach (e.g., TabPFN over XGBoost or GPT‑2’s cache over fine‑tuning) with controlled error rates.
arXiv:2609.35890v1 Announce Type: new Abstract: Sparse-autoencoder decomposability and concentrated feature attribution are increasingly treated as evidence that a model's computation is easier to re...
The paper introduces Unmerge, an efficient machine unlearning algorithm that treats unlearning as the inverse of task arithmetic. By representing the forget component as a low‑rank basis at each layer, Unmerge optimizes three goals—matching the merged vector, suppressing leakage, and bounding correction size—to limit forget leakage and retain damage. Experiments on ResNet‑50, ViT‑S/16, and Llama‑3.2‑3B show significant performance gains over existing methods while maintaining privacy and feature‑distribution fidelity.