ResearchDatasetsLLaMA ChatbotFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsAdversarial Toxicity RefusalLLaMA-2 Chatbot Specialized category59.8Refusal Rate (RTR)3Adversarial Toxicity RefusalLLaMA-2 Chatbot Offensive category47.7RTR3