{
 "about": "Lab 2: the paper's three benchmarks, unedited model vs the released edit, same settings as the paper (thinking on, 32,768-token cap).",
 "runs": [
  {"name": "bench/base/math500", "model": "base", "data": "math500", "n": 4, "shards": 4, "est": 62},
  {"name": "bench/base/gsm8k", "model": "base", "data": "gsm8k", "n": 2, "shards": 5, "est": 77},
  {"name": "bench/base/ifeval", "model": "base", "data": "ifeval", "n": 2, "shards": 2, "est": 27},
  {"name": "bench/p50c/math500", "model": "qwen35-p50c", "data": "math500", "n": 4, "shards": 3, "est": 44},
  {"name": "bench/p50c/gsm8k", "model": "qwen35-p50c", "data": "gsm8k", "n": 2, "shards": 2, "est": 32},
  {"name": "bench/p50c/ifeval", "model": "qwen35-p50c", "data": "ifeval", "n": 2, "shards": 1, "est": 5}
 ]
}