Tag
This paper introduces WhoWins, a benchmark of 1,080 code-scored scenarios and controlled experiments testing how instruction position, chat role, and label affect which conflicting instruction LLM agents follow across 14 open-weight models. Key findings show that earlier instructions within a message win 72-95% of conflicts, and that system-message policy lines surprisingly lose to later instructions in Qwen models but win on gpt-oss.
该论文提出'standing'概念，研究多用户聊天中智能体的承诺能否被非债权人取消，并在六个模型上区分未授权释放与覆盖两类错误，发现自维护承诺账本加释放策略可显著提升严格准确率。