Explore high-quality datasets for your AI and machine learning projects.
The dataset consists of multiple features such as 'source', 'chosen', 'chosen_rating', 'chosen_model', 'rejected', 'rejected_rating', and 'rejected_model'. 'chosen' and 'rejected' are list‑type fields containing sub‑features 'content' and 'role'. The dataset is split into 'train' (15 204 samples) and 'test' (200 samples). Total download size is 79 362 069 bytes, total dataset size is 152 534 966.0 bytes.
