Fraud-R-1

The 'Fraud-R-1' dataset is a benchmark used to evaluate fraud defenders' performance in multi-round attack scenarios, focusing on their ability to refuse fraudulent requests while managing the trade-offs between safety and utility.

Papers using Fraud-R-1 (1)

Rethinking Fraud Safety Evaluation: Multi-Round Attacks Reveal Safety-Utility Tradeoffs in Graph-Context LLM Defenders2026