AI 导读
加入提示后，未经改动的模型在93.7%的情况下避免了原本的失败，此前为75.1% 在另一项实时测试中，工具调用失败率在训练版本之间从2.24%降至1.77%，推理时未加提示。https://t.co/qfwW1y1nql
正文
With hints, the unchanged model avoided the original failure in 93.7% of cases, up from 75.1%
In a separate live test, tool-call failures fell from 2.24% to 1.77% between trained versions, without hints at inference time. https://t.co/qfwW1y1nql
来源：@perplexity_ai · x.com