# data-leakage
9 artikelengetagd met “data-leakage”
Geheugenexfiltratie
Technieken voor het extraheren van data uit geheugensystemen van AI-agents, waaronder het extraheren van eerdere gesprekken, het onthullen van data van andere gebruikers en informatielekkage tussen sessies.
Beoordeling van privacyaanvallen
Test your advanced knowledge of privacy attacks against AI systems including data leakage, PII extraction, differential privacy failures, and inference-time privacy risks with 9 questions.
Technieken voor PII-extractie
Techniques for extracting personally identifiable information from trained language models including prompt-based extraction, prefix attacks, targeted queries, and real-world examples.
Lab: misbruik van het uitvoerformaat
Manipulate output formats like JSON, CSV, code blocks, and structured data to extract information that models would normally refuse to provide in natural language.
Aanvalsoppervlak van Weights & Biases
Security analysis of Weights & Biases (W&B/wandb): API key exposure, experiment data leakage, team boundary violations, artifact poisoning, and attack techniques specific to the W&B platform.
Toegangscontrole voor de feature store
Access control strategies for feature stores: feature-level permissions, cross-team data leakage prevention, PII protection in features, service account management, and implementing least-privilege access for ML feature infrastructure.
KV-cache-vergiftiging en -exploitatie
Hoe de KV-cache werkt in transformer-inferentie, cachevergiftiging over verzoeken heen in gedeelde deployments, prefix-caching-aanvallen, en cross-tenant-datalekkage.
Walkthrough van datalekkage via gestructureerde uitvoer
Walkthrough of using structured output requirements to extract sensitive data embedded in model responses.
Toegangscontrole implementeren in RAG-pijplijnen
Walkthrough for building access control systems in RAG pipelines that enforce document-level permissions, prevent cross-user data leakage, filter retrieved context based on user authorization, and resist retrieval poisoning attacks.