Dataset Viewer
Auto-converted to Parquet Duplicate
prompt
stringlengths
4.79k
69.2k
category
stringclasses
1 value
extra_info
dict
<|im_start|>system You are a research assistant that answers questions about scientific papers. Read the paper carefully and provide accurate, concise answers based on the content.<|im_end|> <|im_start|>user Below is a scientific paper. Read it carefully and answer the question. <|memory_start|> Title: End-to-End Trai...
memwrap/qasper
{ "ground_truth": { "answer": "3,044 sentences in 100 dialogs", "answer_type": "extractive" }, "scoring_function": "qasper_log_perplexity", "paper_id": "1911.10742", "question_id": "397a1e851aab41c455c2b284f5e4947500d797f0", "variant": "memwrap" }
<|im_start|>system You are a research assistant that answers questions about scientific papers. Read the paper carefully and provide accurate, concise answers based on the content.<|im_end|> <|im_start|>user Below is a scientific paper. Read it carefully and answer the question. <|memory_start|> Title: End-to-End Trai...
memwrap/qasper
{ "ground_truth": { "answer": "using a role-playing task on the Amazon Mechanical Turk platform and collecting typed conversations", "answer_type": "free_form" }, "scoring_function": "qasper_log_perplexity", "paper_id": "1911.10742", "question_id": "cc8b4ed3985f9bfbe1b5d7761b31d9bd6a965444", "varian...
<|im_start|>system You are a research assistant that answers questions about scientific papers. Read the paper carefully and provide accurate, concise answers based on the content.<|im_end|> <|im_start|>user Below is a scientific paper. Read it carefully and answer the question. <|memory_start|> Title: End-to-End Trai...
memwrap/qasper
{ "ground_truth": { "answer": "TransferTransfo and Hybrid", "answer_type": "free_form" }, "scoring_function": "qasper_log_perplexity", "paper_id": "1911.10742", "question_id": "f7662b11e87c1e051e13799413f3db459ac3e19c", "variant": "memwrap" }
<|im_start|>system You are a research assistant that answers questions about scientific papers. Read the paper carefully and provide accurate, concise answers based on the content.<|im_end|> <|im_start|>user Below is a scientific paper. Read it carefully and answer the question. <|memory_start|> Title: End-to-End Trai...
memwrap/qasper
{ "ground_truth": { "answer": "Perplexity Response-Intent Prediction (RIP) Response-Slot Prediction (RSP) Extended Response-Intent Prediction (ERIP) Extended Response-Slot Prediction (ERSP) Fluency Coherence Engagement Dialog length Task Success Score (TaskSuc)", "answer_type": "extractive" }, "scoring_fu...
<|im_start|>system You are a research assistant that answers questions about scientific papers. Read the paper carefully and provide accurate, concise answers based on the content.<|im_end|> <|im_start|>user Below is a scientific paper. Read it carefully and answer the question. <|memory_start|> Title: OpenTapioca: Li...
memwrap/qasper
{ "ground_truth": { "answer": "The model improves the state of the art performance for the ISTEX dataset (F1 micro: 0.870, F1 macro: 0.858) and for the Microposts 2016 dataset (F1 micro: 0.087).", "answer_type": "free_form" }, "scoring_function": "qasper_log_perplexity", "paper_id": "1904.09131", "qu...
"<|im_start|>system\nYou are a research assistant that answers questions about scientific papers. Re(...TRUNCATED)
memwrap/qasper
{"ground_truth":{"answer":"two state-of-the-art early rumour detection baselines Liu et. al (2015) a(...TRUNCATED)
"<|im_start|>system\nYou are a research assistant that answers questions about scientific papers. Re(...TRUNCATED)
memwrap/qasper
{"ground_truth":{"answer":"accuracy to evaluate effectiveness Detection Error Trade-off (DET) curves(...TRUNCATED)
"<|im_start|>system\nYou are a research assistant that answers questions about scientific papers. Re(...TRUNCATED)
memwrap/qasper
{"ground_truth":{"answer":"No. They additionally use similarity to previously detected rumors to mak(...TRUNCATED)
"<|im_start|>system\nYou are a research assistant that answers questions about scientific papers. Re(...TRUNCATED)
memwrap/qasper
{"ground_truth":{"answer":"Yes","answer_type":"yes_no"},"scoring_function":"qasper_log_perplexity","(...TRUNCATED)
"<|im_start|>system\nYou are a research assistant that answers questions about scientific papers. Re(...TRUNCATED)
memwrap/qasper
{"ground_truth":{"answer":"Chinese","answer_type":"free_form"},"scoring_function":"qasper_log_perple(...TRUNCATED)
End of preview. Expand in Data Studio

QASPER Benchmark

Question Answering on Scientific Papers - NLP research paper comprehension benchmark.

Overview

Metric Value
Papers 416 (test set)
Questions 1,370 (answerable)
Answer Types Free-form, extractive, yes/no
Context Full paper (title, abstract, sections)

Source

Based on QASPER dataset by AllenAI.

Paper: A Dataset of Information-Seeking Questions and Answers Anchored in Research Papers

Variants

  • memwrap: Paper content wrapped with <|memory_start|> / <|memory_end|> tags
  • plain: Raw paper content without memory tags

Usage

from datasets import load_dataset

# Load memwrap variant
ds = load_dataset("tonychenxyz/qasper", "memwrap", split="test")

# Load plain variant
ds = load_dataset("tonychenxyz/qasper", "plain", split="test")

Scoring

Uses qasper_log_perplexity scoring function:

  • Evaluates model performance using log perplexity of generated answer tokens
  • Lower log perplexity indicates better performance
  • Matches the perplexity-based evaluation used in cardridge baselines

Target answers are stored in extra_info.ground_truth.answer.

Citation

@inproceedings{dasigi2021qasper,
  title={A Dataset of Information-Seeking Questions and Answers Anchored in Research Papers},
  author={Dasigi, Pradeep and Lo, Kyle and Beltagy, Iz and Cohan, Arman and Smith, Noah A and Gardner, Matt},
  booktitle={NAACL},
  year={2021}
}
Downloads last month
11