{
 "entries": [
  {
   "benchmark_version": "cain2027.e79.collective_benchmark.v1",
   "commit": "67a0c74fc6ecb05f45f0fdf4b28f100767822ee0 (run manifest); artifacts committed in f4449aa7913623bc034fa4434ca96c2a2919a889",
   "dataset": "100 synthetic multi-domain arithmetic tasks (+20 replication)",
   "evolution": 79,
   "experiment_id": "e79-v1-first-run",
   "headline": {
    "calls": 1186,
    "four_agents": "29/100",
    "grade": "B-",
    "single_direct": "1/100",
    "tokens": 115622
   },
   "model": "Qwen/Qwen2.5-0.5B-Instruct",
   "model_revision": "weight file sha256 fdf756fa7fcbe7404d5c60e26bff1a0c8b8aa1f72ced49e7dd0210fe288fb7fe",
   "publication_status": "PUBLISHED on three homepages (labelled LIVE) and the Hugging Face Space",
   "raw_artifacts": {
    "benchmark.json": "ba02befd1548e1fb6f7bd85769c25445e2cca20c289e9995b416e04941013c97",
    "e79_master_summary.json": "023f690dfeb62e363eea26810adeeaefe10dbf02e18df3cf25649a59072077b3",
    "evidence.json": "b0500b943d2450d1ca2d389c40867d06ce8351f4423dad17e580bf853a617ff1"
   },
   "seeds": "see artifacts/e79/v1/benchmark.json",
   "superseded_by": "e79-v2-accepted",
   "verification": "22/22",
   "verifier": "scripts/evolution79/verify_e79_offline.py"
  },
  {
   "benchmark_version": "cain2027.e79.collective_benchmark.v2",
   "commit": "f4449aa7913623bc034fa4434ca96c2a2919a889",
   "dataset": "104 tasks in 8 domains, seed 79; 32-task subset for collaboration arms; replication seed 791 (14 questions overlap the main set)",
   "evolution": 79,
   "experiment_id": "e79-v2-accepted",
   "headline": {
    "calls": 986,
    "four_agents": "8/32",
    "grade": "C- (composite uses a stale memory value)",
    "single": "11/32",
    "tokens": 167172
   },
   "model": "Qwen/Qwen2.5-0.5B-Instruct",
   "model_revision": "weight file sha256 fdf756fa7fcbe7404d5c60e26bff1a0c8b8aa1f72ced49e7dd0210fe288fb7fe",
   "publication_status": "PUBLISHED on three homepages and clawx evidence directory",
   "raw_artifacts": {
    "bench/model_calls.jsonl": "b8751fabe372af5769164f11f7883783375c5a4be7e3d7a8b1deee6733662612",
    "e79_v2_master_summary.json": "65ddc2b610e424460249d2e87a3ba9fbb5504427bfaf2753c0e2f4bf17dcedf2",
    "evidence_v2.json": "7cf6537b6a26288530b9999dd8bb97538fb4f956107f2c53a10beb3c309db41c"
   },
   "seeds": [
    79,
    791
   ],
   "supersedes": "e79-v1-first-run",
   "verification": "144/144 (reproduced in E80: worktree, clean room, second host)",
   "verifier": "scripts/evolution79/verify_e79_v2_offline.py"
  },
  {
   "benchmark_version": "cain2027.e80.collective_benchmark.v1",
   "commit": "set at commit time; baseline f4449aa7913623bc034fa4434ca96c2a2919a889",
   "dataset": {
    "replication": {
     "families": [
      "code_trace",
      "label_code_review",
      "label_logs",
      "label_notes",
      "schedule",
      "table",
      "word_math"
     ],
     "seed": 8004,
     "sha256": "29c2cdbfccc4c539dfc541f2d2808645a237414cc0cf5e1c0aaee7e87960fd47",
     "tasks": 24
    },
    "test": {
     "families": [
      "code_trace",
      "label_code_review",
      "label_logs",
      "label_notes",
      "schedule",
      "table",
      "word_math"
     ],
     "seed": 8003,
     "sha256": "357de046001a4bb242821d4097d396ad899b7d58dbce400b3887ad419a90f6be",
     "tasks": 48
    },
    "train": {
     "families": [
      "code_trace",
      "label_code_review",
      "label_logs",
      "label_notes",
      "schedule",
      "table",
      "word_math"
     ],
     "seed": 8001,
     "sha256": "79e0e25330faa73eb95ce676c7a33428a89dad45dd55e16ac9d06d9f2b25798c",
     "tasks": 40
    }
   },
   "evolution": 80,
   "experiment_id": "e80-adaptive-routing-and-coordination",
   "model": {
    "qwen2.5-0.5b-instruct": "Qwen/Qwen2.5-0.5B-Instruct",
    "qwen2.5-1.5b-instruct": "Qwen/Qwen2.5-1.5B-Instruct",
    "smollm2-135m-instruct": "HuggingFaceTB/SmolLM2-135M-Instruct"
   },
   "model_revision": {
    "qwen2.5-0.5b-instruct": "fdf756fa7fcbe7404d5c60e26bff1a0c8b8aa1f72ced49e7dd0210fe288fb7fe",
    "qwen2.5-1.5b-instruct": "dd924a11b4c220f385b51ffa522daea7c9f3d850e31b162bb5661df483c6d3ee",
    "smollm2-135m-instruct": "5af571cbf074e6d21a03528d2330792e532ca608f24ac70a143f6b369968ab8c"
   },
   "publication_status": "NOT PUBLISHED (development branch only)",
   "raw_artifacts": {
    "collective_benchmark.json": "7b9715e0064c0aefe249a406472c932bef6c37135ea457417d64965233ffcece",
    "model_calls.jsonl": "fc4bc05efe830a6b52151b01c1a5f56af68a806bb6a2eb7c6f797ed08d47718f",
    "preregistration.json": "5d430d817f8f9c77e4ddd5ce519b57f2801c033175415be03df4368f3b8b133b",
    "routing_policy.json": "dd010dd83e2b02122573f9e84a7f98396252bc1c2419bb28d168ee2c0ad46372"
   },
   "seeds": {
    "replication": 8004,
    "test": 8003,
    "train": 8001
   },
   "supersedes": null,
   "verification": "see offline_verification.json",
   "verifier": "scripts/evolution80/verify_e80_offline.py"
  }
 ],
 "generated_utc": 1794787210.201,
 "schema": "cain2027.e80.canonical_evidence_index.v1"
}