BFCL

Emerging

1papers using it

2026first seen

The 'BFCL' dataset/benchmark is used to evaluate the vulnerabilities of function calling in Large Language Models (LLMs) by examining the effectiveness of function hijacking attacks and other related security threats.

🔎 Find this dataset

Papers using BFCL (1)

Breaking MCP with Function Hijacking Attacks: Novel Threats for Function Calling and Agentic Models2026