用MUD游戏测LLM社交能力
推荐指数 55.0 NO. 017 · 2026.07.23
发布2026/07/22Score56Comments28
为什么值得看
CrucibleBench把大模型放进持续运行的文字MUD世界，通过50轮交互中的信任积累、记忆回溯和错误追踪来评分。这比静态benchmark更能暴露模型在复杂社交场景中的真实缺陷，成本只要99美元。
编辑判断
当前LLM评测的主流是静态QA和短期任务，但真实应用场景需要模型在多轮交互中维持一致性、修复错误、管理社交关系——这些恰恰是现有benchmark的盲区。CrucibleBench的聪明之处在于用MUD的有限命令空间天然约束了幻觉，NPC的信任状态提供了可量化的社交反馈，这比动辄需要复杂多智能体框架或3D仿真的方案轻量得多。
对比SWE-bench测代码、HumanEval测算法，这个方向补的是"长期社交智能"的评测缺口。做Agent团队的人值得关注：如果你的产品需要模型在多轮对话中维护用户关系或修复信任，现有的短期评测可能严重高估了模型能力。代码未明确是否开源，但99美元的复现成本足够低，建议直接跑一轮看自己的模型在哪个turn开始崩。
社区反馈
意见分歧 29 条评论
核心争论：MUD能否借LLM复兴，还是Discord等现代平台已不可逆地取代文本社交游戏
相关内容
A little irrelevant, but I need to vent about something MUD adjacent. I really wish MUDs were still a thing. The text-based roleplaying communities were largely swallowed by Discord, where as before they existed scattered over proboards, jcink, and a few others. Discord, being the imperfect text pla
I agree with you. You've also got more / easier competition. IIRC, MUDs peaked in the mid-late 90s. MMOs were really the death knell, then the explosion of other types of games, mobile games, then to your point things like Discord. One adjacent project we're exploring for the future is if you t
I honestly worry that the relevance of text-based systems is largely tied with literacy and how it interacts with the player's imagination -- something largely in decline, at least in the US.