Sign in to view source links and access this dataset
Description
A July 2024 collection of datasets for generating and evaluating essays. It includes essays with instructions, IELTS essays with human feedback, persuasive essays with human scores, and Change My View (CMV) subreddit posts. The CMV samples use the Qwen2-72B-Instruct model to generate instructions for essays supporting and opposing original claims.
Use Cases
Training instruction-following models based on the essay-with-instructions data.
Evaluating essay quality using the human feedback scores from IELTS and persuasive essay datasets.
Generating opposing viewpoints for argumentative writing based on the CMV subreddit posts and AI-generated instructions.
Strengths
Includes multiple source datasets (ChristophSchuhmann/essays-with-instructions, chillies/IELTS_essay_human_feedback, persuade_2.0_human_scores_demo_id_github.csv, Change My View) that have been deduplicated.
Utilizes a large language model (Qwen/Qwen2-72B-Instruct) to generate specific instructions for argumentative essays.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Row count is unknown, which may limit suitability assessment.
Last updated 2024-07-12 21:18:56; freshness should be verified.
Provenance
Source
HuggingFace datasets: ChristophSchuhmann/essays-with-instructions, chillies/IELTS_essay_human_feedback, persuade_2.0_human_scores_demo_id_github.csv, and Change My View (CMV) subreddit.
Collection Method
Collection and deduplication of existing datasets; AI-generated instructions for CMV posts.