{
  "id": "benchmark:promptbench",
  "_kind": "Benchmark",
  "_file": "benchmarks/benchmarks/benchmarks-knowledge.yaml",
  "_cluster": "benchmarks",
  "attributes": {
    "displayName": "PromptBench",
    "homepageUrl": "https://github.com/microsoft/promptbench",
    "kind": "model-only",
    "targetsKind": "ModelVersion",
    "description": "PromptBench (Microsoft) measures the robustness of LLMs to prompt-\nlevel adversarial perturbations (character-, word-, sentence-, and\nsemantic-level attacks) across classification, QA, reasoning, and\ntranslation tasks.\n"
  },
  "outgoingEdges": [
    {
      "from": "benchmark:promptbench",
      "to": "skill-area:prompt-engineering",
      "kind": "covers",
      "attributes": {}
    },
    {
      "from": "benchmark:promptbench",
      "to": "domain:ml-ai",
      "kind": "applies_to",
      "attributes": {}
    },
    {
      "from": "benchmark:promptbench",
      "to": "domain:security",
      "kind": "applies_to",
      "attributes": {}
    }
  ],
  "incomingEdges": []
}