README.md

metadata

license: cc-by-4.0
language:
  - en
tags:
  - peer-review
  - scientific-papers
  - nlp
size_categories:
  - 10K<n<100K

BMC Medicine Peer Reviews

This dataset contains peer reviews from BMC, standardized to match the format of the pawin205/PeerRT dataset.

Dataset Description

• Source: BMC
• Standardization: Matches pawin205/PeerRT schema with an additional year attribute.
• Total Records: (See dataset viewer)

Dataset Structure

Each record contains the following attributes:

• relative_rank: Default value (0).
• win_prob: Default value (0.0).
• title: Title of the paper.
• abstract: Abstract of the paper.
• full_text: Full text of the paper (or review text if unavailable).
• review: The peer review text.
• source: Source of the data ('BMC').
• review_src: Original filename of the review.
• criteria_count: Dictionary containing counts of specific criteria in the review (e.g., criticism, praise).
• dimension_scores: Dictionary containing scores for various dimensions.
• overall_score: Computed overall score (reward value).
• thinking_trace: Placeholder (empty).
• prompt: Placeholder (empty).
• prompt_length: Placeholder (0).
• conversations: List of dictionaries representing the conversation (User/Assistant format).
• year: Year of publication (if available).

Usage

from datasets import load_dataset

dataset = load_dataset("JerMa88/BMC_Medicine_Peer_Reviews")
print(dataset['train'][0])