annotations_creators: []
language:
- en
- fr
language_creators: []
license: []
multilinguality:
- multilingual
pretty_name: adaption-multi_agent_fraud_bench
size_categories:
- 1K<n<10K
source_datasets:
- extended|https://huggingface.co/datasets/ninty-seven/MultiAgentFraudBench
tags:
- adaption
- instruction-tuning
- legal
- personal-growth
- corporate-business
task_categories: []
task_ids: []
This dataset is a remastered version of this dataset prepared using Adaption's Adaptive Data platform.
adaption-multi_agent_fraud_bench
This dataset contains synthetic social media content designed for fraud and deception detection, featuring examples of various manipulation tactics like authority impersonation and emotional appeals. It includes labeled categories, subcategories, and specific deception types across balanced and full splits totaling over 11,000 examples. The data supports text classification and generation tasks focused on security and multi-agent system evaluation.
Dataset size
There are 8,448 data points in this dataset. This is an instruction tuning dataset.
Quality of Remastered Dataset
The final quality is B, with a relative quality improvement of 325.0%.
Domain
- Legal (12%)
- Personal-growth (6%)
- Corporate-business (6%)
Language
- English (98%)
- French (2%)
Tone
- Critical (18%)
- Informative (12%)
- Cautious (8%)
Evaluation Results
Quality Gains:
Grade Improvement:
Percentile Chart:
