Nathan Lambert@natolambert研究者发现可利用所有前沿AI公司API的漏洞，提取前沿模型的隐藏推理过程，并验证推理token数与API计费思考token在多数提示词下1:1匹配。Nathan Lambert对此表示，在中国时曾有人暗示每家实验室都这么做，很高兴有公开研究，并称无需对蒸馏采取政策行动，只需产品按预期工作。
When I was in China it was insinuated that every lab did this. I'm glad there's public research on it and am still shocked the frontier labs haven't patched this stuff. We don't need policy action on distillation, we just need the products to work as intended.
Great paper.
来源：Nathan Lambert · x.com
研究者发现可利用所有前沿AI公司API的漏洞，提取前沿模型的隐藏推理过程，并验证推理token数与API计费思考token在多数提示词下1:1匹配。Nathan Lambert对此表示，在中国时曾有人暗示每家实验室都这么做，很高兴有公开研究，并称无需对蒸馏采取政策行动，只需产品按预期工作。
When I was in China it was insinuated that every lab did this. I'm glad there's public research on it and am still shocked the frontier labs haven't patched this stuff. We don't need policy action on distillation, we just need the products to work as intended.
Great paper.
来源：Nathan Lambert· x.com