AI safety | Bardh Prenkajhttps://bardhprenkaj.it/tag/ai-safety/AI safetyWowchemy (https://wowchemy.com)en-usWed, 25 Mar 2026 00:00:00 +0000https://bardhprenkaj.it/media/icon_hu0b7a4cb9992c9ac0e91bd28ffd38dd00_9727_512x512_fill_lanczos_center_3.pngAI safetyhttps://bardhprenkaj.it/tag/ai-safety/Analysing the Safety Pitfalls of Steering Vectorshttps://bardhprenkaj.it/publication/li-2026-steering/Wed, 25 Mar 2026 00:00:00 +0000https://bardhprenkaj.it/publication/li-2026-steering/Injecting Falsehoods: Adversarial Man-in-the-Middle Attacks Undermining Factual Recall in LLMshttps://bardhprenkaj.it/publication/fastowski-2025-falsehoods/Sat, 08 Nov 2025 00:00:00 +0000https://bardhprenkaj.it/publication/fastowski-2025-falsehoods/Is My Data Safe? Predicting Instance-Level Membership Inference Success for White-box and Black-box Attackshttps://bardhprenkaj.it/publication/leemann-2024-membership/Fri, 26 Jul 2024 00:00:00 +0000https://bardhprenkaj.it/publication/leemann-2024-membership/