Yibaii/selfrag_train_data
This is a training data file for Self-RAG that generates outputs to diverse user queries as well as reflection tokens to call the retrieval system adaptively and criticize its own output and retrieved passages. Self-RAG is trained on our 150k diverse instruction-output pairs with interleaving passages and reflection tokens using the standard next-token prediction objective, enabling efficient and stable learning with fine-grained feedback. At inference, we leverage reflection tokens covering… See the full description on the dataset page: https://huggingface.co/datasets/Yibaii/selfrag_train_data.
026
1---2license: mit3task_categories:4- text-generation5language:6- en7size_categories:8- 100K<n<1M9---10 11This is a training data file for [Self-RAG](https://selfrag.github.io/) that generates outputs to diverse user queries as well as reflection tokens to call the retrieval system adaptively and criticize its own output and retrieved passages.12 13Self-RAG is trained on our 150k diverse instruction-output pairs with interleaving passages and reflection tokens using the standard next-token prediction objective, enabling efficient and stable learning with fine-grained feedback.14At inference, we leverage reflection tokens covering diverse aspects of generations to sample the best output aligning users' preferences. See full descriptions in [our paper](https://arxiv.org/abs/2310.11511) and [code](https://github.com/AkariAsai/self-rag).15 16## Citation and contact17If you use this model, please cite our work: 18```19@article{asai2023selfrag,20 author = {Asai, Akari and Wu, Zeqiu and Wang, Yizhong and Sil, Avirup and Hajishirzi, Hannaneh},21 title = {{Self-RAG}: Learning to Retrieve, Generate, and Critique through Self-Reflection},22 year = {2023},23 journal = { arXiv preprint arXiv:2310.11511 },24 URL = {https://arxiv.org/abs/2310.11511}25}26```