{
  "generated": "2026-09-08",
  "benchmark": "hellaswag",
  "results": [
    {
      "system": "BERT-Large",
      "developer": "Google (evaluated by Zellers et al.)",
      "value": 47.3,
      "date": "2019-05-19",
      "source": {
        "url": "https://arxiv.org/abs/1905.07830",
        "kind": "paper",
        "accessed": "2026-09-04"
      },
      "conditions": {
        "notes": "Table 1 of the HellaSwag paper, overall test accuracy; humans 95.6."
      }
    },
    {
      "system": "DeepSeek-V4-Pro-Base",
      "developer": "DeepSeek",
      "value": 88.0,
      "date": "2026-04-22",
      "source": {
        "url": "https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro",
        "kind": "developer-report",
        "accessed": "2026-09-04"
      },
      "conditions": {
        "shots": 0,
        "notes": "DeepSeek-V4 model card, base-model table (HF repo created 2026-04-22). Base model, exact match, zero-shot; not comparable with 10-shot harness numbers."
      }
    }
  ]
}