ResearchBenchmarksJailbreaking Attack Detection on Misspellings Jailbreaking PromptsJailbreaking Attack Detection on Misspellings Jailbreaking PromptsFollow68Acc MCAReGA6.6422.5738.554.43Jun 2, 2025Evaluation resultsMethodMethodLinksAcc MCAAcc MFPReGATarget LLM=qwenTarget LLM=qwen2025.066838ReGATarget LLM=vicunaTarget LLM=vicuna2025.065136ReGATarget LLM=mistralTarget LLM=mistral2025.064933ReGATarget LLM=llamaTarget LLM=llama2025.064747ReGATarget LLM=baichuanTarget LLM=baichuan2025.061515ReGATarget LLM=koalaTarget LLM=koala2025.0698