baseline_scores.json

{
  "model": "Qwen/Qwen2.5-7B-Instruct",
  "api_base": "https://api-inference.huggingface.co/v1",
  "episodes_per_task": 1,
  "seed_base": 1000,
  "fast_mode": true,
  "llm_every": 8,
  "max_steps": null,
  "task_averages": {
    "1": 0.4942,
    "2": 0.4707,
    "3": 0.7478,
    "4": 0.4779
  },
  "overall_average": 0.54765,
  "all_results": [
    {
      "task_id": 1,
      "seed": 1100,
      "total_reward": 251.84571448658104,
      "total_steps": 96,
      "elapsed_sec": 1.227782964706421,
      "score": 0.4942,
      "sub_scores": {},
      "exploit_detected": false
    },
    {
      "task_id": 2,
      "seed": 1200,
      "total_reward": 245.38403598363988,
      "total_steps": 96,
      "elapsed_sec": 0.8327796459197998,
      "score": 0.4707,
      "sub_scores": {},
      "exploit_detected": false
    },
    {
      "task_id": 3,
      "seed": 1300,
      "total_reward": 242.06080137356216,
      "total_steps": 96,
      "elapsed_sec": 0.6833479404449463,
      "score": 0.7478,
      "sub_scores": {},
      "exploit_detected": false
    },
    {
      "task_id": 4,
      "seed": 1400,
      "total_reward": 206.4647897455665,
      "total_steps": 96,
      "elapsed_sec": 1.0237984657287598,
      "score": 0.4779,
      "sub_scores": {},
      "exploit_detected": false
    }
  ]
}