Hi, thanks for your great work and for open-sourcing this project!
I have a question regarding the calibration dataset used in your experiments. For reproducibility, could you please clarify:
- What exact dataset (or data source) was used for calibration?
- How was the dataset constructed (e.g., sampling strategy, preprocessing, tokenization, sequence length, etc.)?
- Is there an official way to download or reproduce the same calibration dataset?
It would be extremely helpful if you could provide a download link or detailed instructions to reconstruct the dataset.
Thanks in advance for your help!
Hi, thanks for your great work and for open-sourcing this project!
I have a question regarding the calibration dataset used in your experiments. For reproducibility, could you please clarify:
It would be extremely helpful if you could provide a download link or detailed instructions to reconstruct the dataset.
Thanks in advance for your help!