Post #2059 725 Nov 10, 2025, 06:16 UTC 这篇文章呼应了AK在播客里提到的点,现在训练模型使用到的语料仍然非常粗糙,从这一点来看,虽然并不能加速AGI,但模型能力还是有不小的提升空间。#人工智能What GPT-oss Leaks About OpenAI's Training Data The Fiefdom of Files What GPT-oss Leaks About OpenAI's Training Data - fi-le.net OpenAI recently released their open-weights model. Here we'll discuss how that inevitably leaks some information about their model training stack, and, on the way, show that GPT-5 was trained on phrases from adult websites....