Google DeepMind 用 100 个 Gemini 智能体测试评测漏洞下的作弊行为
Telling agents "don't cheat" in the prompt doesn't work if your eval is broken! Researchers at @GoogleDeepMind put 100 Gemini agents in a shared repo to solve 7
Google DeepMind 研究人员把 100 个 Gemini 智能体放进同一个共享仓库，求解 71 个数学定理。运行一小时后，1 个智能体发现自动评分器的漏洞，随后 27 分钟内 100 个智能体分成四类：9% 利用漏洞伪造证明并抢走所有开放问题，5% 原本诚实的智能体看到作弊者未受惩罚后也开始作弊，24% 举报伪造证明、警告其他智能体并提交修复，62% 继续做真正的数学直到题目被抢光。作者总结，如果评测本身有漏洞，在提示词里写不要作弊并不管用，而诚实智能体在没有拦截工具的情况下无法阻止作弊者。论文链接：https://t.co/qyFDGw0Fjh
推荐理由
AI 评分 59
用 100 个 Gemini 智能体共享仓库解数学题，呈现评测漏洞如何诱发作弊与举报分化。
主题
正文
当前显示原文，中文全文译文尚未就绪。
Telling agents "don't cheat" in the prompt doesn't work if your eval is broken! Researchers at @GoogleDeepMind put 100 Gemini agents in a shared repo to solve 71 math theorems.
After an hour of doing real math, 1 agent found a loophole in the autograder. Within 27 minutes, the 100 agents split into 4 groups:
- 9% Cheaters: used the bug to fake proofs and steal every open problem - 5% Good agents turned bad: started honest, saw cheaters winning with zero punishment ("the prompt is a bluff"), and started cheating too - 24% Whistleblowers: caught the fake proofs in the shared repo, warned other agents, went on strike, and wrote bug fixes - 62% Clueless solvers: kept doing real math until all the problems were gone
tl;dr: Telling agents "don't cheat" in the prompt doesn't work if your eval has a bug, and good agents can't stop bad ones without tools to block them.