txy

@icedpanda Thank you very much for your efforts in reproducing the results. We will try our best to ensure that the results is consistent with the paper. However, due to...

Thanks for your feedback. However, there is a problem with Baidunetdisk. If we update the data on Baidunetdisk, the user must only download it manually. But onedrive allows users to...

Thank you for your feedback. I've fixed the bug now, you can try again.

Thanks for your feedback. @Oran-Ac 1. The padding labels are set to **-100** according to your suggestion. 2. Due to the limitation of gpu memory, we cuf off the context...

Thank you for your attention! We measure the model's preference for the stereotypical sentence using the perplexity of both sentences in a zero-shot setting. "sent_more_ppl_score" represents the perplexity of the...

Thank you for your attention. The running time depends on the device you are using.

We are going to update it soon. Please use python 3.7 for now.

I wonder whether this issue has been resolved

Is the path you provided, https://olmo-checkpoints.org/ai2-llm/olmo-medium/wvc30anm/train_data/global_indices.npy, correct?