Buyun-Liang/REALISTA

[ICML 2026] REALISTA: Realistic Latent Adversarial Attacks that Elicit LLM Hallucinations

语言：Jupyter Notebook · GitHub 星标：30 · Fork：0

话题：adversarial-attacks · large-language-models · llm-hallucination · llm-safety

在 GitHub 上查看原仓库