1
Epistemic Bias Injection: Manipulating LLM Opinion via Selective Context Retrieval
用几何度量揭示检索上下文如何操纵大模型立场，轻量攻击原型曝光AI偏见注入风险。
arXiv:2512.00804v3 Announce Type: replace-cross Abstract: When answering user queries, LLMs often retrieve knowledge from external sources stored in r…