The paper introduces the task of Scientific Claim Unlearning and presents a new benchmark, SciUnlearn, to evaluate it. It highlights that language models trained on static scientific corpora risk disseminating outdated or retracted claims as scientific knowledge evolves. Current machine unlearning methods fail to effectively remove claim-level knowledge, often only suppressing it superficially, underscoring the need for specialized techniques for structured knowledge removal.
By Snigdha Paul, Manasi Patwardhan, Arman Cohan
arXiv:2606.17467v3 Announce Type: replace-cross
Abstract: Prompt injection defenses evaluated on synthetic benchmarks do not generalize to real enterprise documents, which are longer, denser, and int...
By Aaditya Pai
arXiv:2608. 10509v1 Announce Type: new Abstract: Shared memory helps language-model agents reuse information across long workflows, yet relevant evidence may not be admissible for a particular agent or action.
By Yiqi Wang, Zihao Yan, Jiaqi Zhang, Zhangkai Wu, Mingkai Zheng, Zequn Sun, Yanming Zhu, Taotao Cai
Shared memory helps language-model agents reuse information across long workflows, yet relevant evidence may not be admissible for a particular agent or action. Because restrictions propagate through derivations, summaries can conceal private, poisoned, untrusted, or revoked sources, enabling unauthorized reads or unsafe actions.
arXiv:2608.21230v1 Announce Type: cross
Abstract: Persistent memory makes false information durable: once a false statement is stored, it can be retrieved into future sessions that match it. We measu...
By Arulnidhi Karunanidhi
arXiv:2607. 09236v1 Announce Type: new Abstract: Machine unlearning in LLMs is the targeted removal of specific knowledge while preserving all other capabilities, critical for privacy and safety.
By Amit Peleg, Naman Deep Singh, Naama Pearl, Bibhabasu Mohapatra, Matthias Hein
arXiv:2601. 22601v3 Announce Type: replace Abstract: Federated unlearning (FU) aims to erase knowledge from a global model.
By Wentai Wu, Hanwei Tan, Yijun Quan, Haixia Peng, Ligang He, Bin Yang, C. L. Philip Chen
The paper introduces a provenance-guided incremental learning framework that handles rule-induced concept shift, where target definitions are explicitly revised and previously stored instances receive new semantic labels. By compiling concept changes into structured rule deltas, tracing affected records through historical provenance, and selectively re-evaluating only a localized candidate region, the method automatically relabels executable revisions, manages ambiguous cases with selective supervision, and repairs predictors incrementally. Evaluation on the RuleShift-Bench benchmark—covering financial, demographic, cybersecurity, and graph-structured data—shows 92.3% accuracy and 90.2% Macro‑F1, reprocessing only 14.7% of the historical collection and achieving an average update latency of 179 s versus 993 s for full relabeling and retraining.
By Ismail Lamaakal
The paper introduces GRAPHSU, a graph‑guided selective unlearning method for language models that expands deletion beyond explicitly identified forget seeds. By constructing a weighted support‑route graph and propagating deletion pressure, GRAPHSU applies graded forgetting to high‑risk neighboring examples. Experiments on the TOFU and PISTOL benchmarks with GPT‑2 Medium and Llama‑3.2‑3B‑Instruct show that GRAPHSU achieves the lowest utility‑feasible soft leakage, reducing leakage by up to 49.5 percentage points compared to a seed‑only baseline.
By Waqas Khan, Tabinda Sarwar, Jingyue Cong, Xun Yi, Estrid He
arXiv:2605. 07482v2 Announce Type: replace Abstract: Machine unlearning for large language models (LLMs) aims to selectively remove memorized content such as private data, copyrighted text, or hazardous knowledge, without costly full retraining.
By Zizhao Hu, Ameya Godbole, Johnny Tian-Zheng Wei, Mohammad Rostami, Jesse Thomason, Robin Jia
arXiv:2608. 11022v1 Announce Type: cross Abstract: Model Cards and Data Cards have demonstrated the value of structured, human-readable documentation for machine learning artifacts, capturing their context, parameters, limitations, and intended use.
By Nicola Giuseppe Marchioro, Gabriele Padovani, Amal Gueroudji, Rafael Ferreira da Silva, Wesley Brewer, Valentine Anantharaj, Sandro Fiore, Renan Souza
arXiv:2604. 01904v3 Announce Type: replace-cross Abstract: Post-hoc unauthorized-training data detection for large language models (LLMs) typically assumes a query-with-originals regime: rights holders query a target LLM with raw proprietary data and assess whether the model assigns them stronger memorization-based detection signals, e.
By Muxing Li, Zesheng Ye, Sharon Li, Feng Liu