Request for Updating the Link of the Aligned Model Weights

[Raytsang123]

Hi authors,

This is a great work. It seems that in the second version of the paper, the authors additionally conducted alignment on the industry-level instruction-tuned models (e.g., Llama-3-8B-Instruct); while in the first version, the authors primarily conducted in-lab instruction tuning on the base models (e.g., Llama-3-8B). Given that the model checkpoints released in this code repository are mainly connected to the base models, could the authors further release the weights of the hardened industry-level instruction-tuned models? It would be benificial to the research community. Thanks a lot!