As this is a public benchmark, and model companies will scrape web to train their models. Do you have a string or disclaimer to avoid this dataset to be excluded from training to be able to accurately measure the new model's capabilities?
· Sign up or log in to comment
