Dataset · allenai
Questions and answers on NLP research papers
allenai/qasper
1,585 NLP papers and 5,049 reader questions with practitioner answers for QA training and evaluation.
What's inside
The dataset contains 1,585 NLP research papers and 5,049 information-seeking questions asked by readers. Answers were provided by individual NLP practitioners. The language is English, and the modality is text.
What you can do with it
- Training QA models on scientific texts
- Evaluating answer retrieval in NLP papers
- SFT of an assistant for working with scientific literature
- Tasks
- Question answering
- Languages
- English
- Size
- 1K–10K rows
- Downloads this month
- 13.1K
- Likes
- 129
Models trained on it
- mradermacher/mpt-30b-instruct-i1-GGUF 2.3K
- mradermacher/mpt-30b-instruct-GGUF 1.2K
- xxparthparekhxx/openjev-flashText classification 65
- parsecai/ragzillaText classification 56
- anmol-unitmole/longformer-qasper-document-qaQuestion answering 38
- whosouravsharma/paper-qa-loraText generation 31
- gouneji/all-MiniLM-L6-v2-qasperSentence similarity 27
- gouneji/ms-marco-MiniLM-L6-v2-qasperText ranking 27
How to load it
from datasets import load_dataset
ds = load_dataset("allenai/qasper")