arXiv AI August 11, 2026 By Michael Levit, Josh Ledgard, Haoyu Dong, Vishwas Suryanarayanan, Eyal Kolman, Sharon Tan, Qiang Gan, Vishal Chowdhary

Privacy-Preserving Data Drift Detection and Recovery for Large-Scale LLM Applications via Proxy Representations

Read the original on arXiv AI →

arXiv:2608. 08245v1 Announce Type: cross Abstract: LLM applications deployed at scale face a fundamental challenge: privacy constraints prevent direct inspection of user interactions, making it difficult to obtain any representative evaluation dataset or to track the ongoing evolution of production traffic.

Summary generated by The Flow from the publisher's feed. The full article lives at arXiv AI.

arXiv Machine Learning

Jun 9

A Framework for Evaluating and Benchmarking Concept Drift Detection Methods

arXiv:2606. 07789v1 Announce Type: new Abstract: Data stream mining is fundamentally challenged by concept drift, where distributional changes can degrade model performance.

By Vitor Cerqueira, Heitor Murilo Gomes, Marco Heyden, Bernhard Pfahringer, Albert Bifet

benchmarks

arXiv AI

2d ago

Proxy-Validated LLM UX Micro-Simulations: An Artifact-First Protocol for Early-Stage Decision Support

arXiv:2608. 13563v1 Announce Type: cross Abstract: Early-stage teams often lack users, time, and budget to run repeated UX studies, yet still need decision-oriented signals to iterate safely.

By Alexandre Cristov\~ao Maiorano

llms rag agents safety

arXiv AI

Jul 22

Phantoms and Disclosures: A Statistical Framework for Auditing Privacy in Synthetic Data

arXiv:2606. 16952v2 Announce Type: replace-cross Abstract: The rapid adoption of generative AI and Large Language Models (LLMs) has spurred interest in synthetic data as a privacy-preserving alternative to sensitive real-world datasets.

By Kareem Amin, Rudrajit Das, Alessandro Epasto, Adel Javanmard, Dennis Kraft, M\'onica Ribero, Sergei Vassilvitskii

llms safety

arXiv Machine Learning

Jun 10

LLM-as-a-Discriminator: When Synthetic Tables Still Look Real

arXiv:2606. 09865v1 Announce Type: new Abstract: Privacy and data sharing are often in tension.

By Manel Slokom, Malek Slokom, Thierno Kante

llms safety

arXiv Machine Learning

Jul 20

Do Generative Models Keep Time? A Time-Aware Evaluation of Synthetic Sequential Tabular Data

arXiv:2607. 15606v1 Announce Type: new Abstract: Synthetic sequential tabular data are increasingly used for privacy-preserving data sharing, yet a generator can reproduce every marginal and every foreign-key relationship while emitting timestamps that run backwards or repeat, and while sending entities along paths that no real entity followed.

By Kiwan Kwon, Kangmin Kim, Hojin Lee, Yeseong Jung, Hyeongwoo Kong, Vamsi K. Potluru, Saerom Park, Yongjae Lee

safety

arXiv Machine Learning

Aug 12

Seq2Synth: Benchmarking Temporal Fidelity in Synthetic Sequential Tabular Data

arXiv:2607. 15606v2 Announce Type: replace Abstract: Synthetic sequential tabular data are increasingly used for privacy-preserving data sharing and data-driven research, but evaluating their fidelity remains difficult because temporal structure is easily lost under conventional tabular metrics.

By Kiwan Kwon, Kangmin Kim, Hojin Lee, Yeseong Jung, Hyeongwoo Kong, Vamsi K. Potluru, Saerom Park, Yongjae Lee

benchmarks safety

Related stories

A Framework for Evaluating and Benchmarking Concept Drift Detection Methods

Proxy-Validated LLM UX Micro-Simulations: An Artifact-First Protocol for Early-Stage Decision Support

Phantoms and Disclosures: A Statistical Framework for Auditing Privacy in Synthetic Data

LLM-as-a-Discriminator: When Synthetic Tables Still Look Real

Do Generative Models Keep Time? A Time-Aware Evaluation of Synthetic Sequential Tabular Data

Seq2Synth: Benchmarking Temporal Fidelity in Synthetic Sequential Tabular Data