2 datasets found

Tags: Backdoor attacks

Filter Results

  • AgentInstruct

    The dataset used in the paper to investigate backdoor threats to LLM-based agents.

  • ToolBench

    The ToolBench dataset is a comprehensive instruction-tuning dataset for tool utilization, encompassing 16,000 tools and 49 categories.
You can also access this registry using the API (see API Docs).