Datasets:
| license: mit | |
| task_categories: | |
| - question-answering | |
| language: | |
| - en | |
| tags: | |
| - HotpotQA | |
| - TriviaQA | |
| - NaturalQuestions | |
| - MuSiQue | |
| - 2WikiMultihopQA | |
| - PAI-2 | |
| size_categories: | |
| - 1K<n<10K | |
| This is a repository with a preprocessed well known benchmarks that were used for PAI-2 (GraphRAG-method from [PAI framework](https://github.com/Dzigen/PersonalAI)) evaluation on QA tasks. More details in our paper: arxiv.org/abs/2605.13481. |