AI的“读心术”：模型的大脑会出卖它的秘密吗？

[CL] Approximating Language Model Training Data from Weights

[Cornell University]

arxiv.org