你会为一个执行LLM代码以进行验证的工具付费吗?

1作者: ZOdex大约 2 个月前原帖
我正在构建一个包装器,查询 GPT-4、Claude 和 Gemini,然后在沙盒中执行它们的代码,以捕捉幻觉。<p>你觉得延迟(30秒)值得这种确定性吗?还是你更喜欢速度?<p>今天我正在为一些人进行手动测试,如果有人想试试的话。
查看原文
m building a wrapper that queries GPT-4, Claude, and Gemini, then executes their code in a sandbox to catch hallucinations.<p>Is the latency (30s) worth the certainty? Or do you prefer speed?<p>I&#x27;m running manual tests for people today if anyone wants to try it.