ELBA-Bench introduces a comprehensive framework for assessing backdoor attack vulnerabilities in large language models where subtle triggers can compromise model behavior.

• Addresses gaps in existing benchmarks with improved coverage of attack methods
• Provides an integrated metric system for evaluating backdoor vulnerabilities
• Creates realistic scenarios that align with practical constraints
• Enables more effective testing and development of defense mechanisms

This research is critical for the security community as it helps identify and mitigate risks in widely-deployed LLMs, preventing potential exploitation through backdoor attacks in real-world applications.

ELBA-Bench: An Efficient Learning Backdoor Attacks Benchmark for Large Language Models