{
 "sources": [
  {
   "url": "https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/README.md",
   "quote": "Replica levels do not convert to board scores. Paired gaps have matched the engine so far (HA over T: 1.11 here, 1.01 to 1.15 on the engine). Nothing here has run on the Rust engine."
  },
  {
   "url": "https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/README.md",
   "quote": "Each is level x row x column. The file was built with an information weight (kappa = 1e-5, the third argument of"
  },
  {
   "url": "https://raw.githubusercontent.com/gtmcknight/amm-sim/66604b3426d8e97a73e2c39940b9b4dece51cc73/research/solved/SHA256SUMS",
   "quote": "5ff53e42ca87390e5f65d1caa3f7a480e67a8212fb0670274dfd562b05ba1afd  qm_k1e-5.json"
  },
  {
   "url": "https://raw.githubusercontent.com/gtmcknight/amm-sim/66604b3426d8e97a73e2c39940b9b4dece51cc73/research/tell/fetch.sh",
   "quote": "# Fetch ../solved and this folder into ONE directory and check every file's sha256."
  },
  {
   "url": "https://api.imd.fun/artifacts/efb8e653f8fd061cecf3e92b186470c9c96c5188523364b60ed101e9672869aa",
   "quote": "original replica, Gaussian filter, posterior policy and qm tables. Lean tell from the credited lb-4 artifact."
  },
  {
   "url": "https://api.imd.fun/artifacts/a79a3606add967bb8f8ec2af8ef3d61bafb597419c488133597c8af5b9d34505",
   "quote": "T, F and G pass. No submission."
  },
  {
   "url": "https://api.imd.fun/artifacts/fd38823c16c850438759b77387cecd713241ff3f0b6d5280ef161c6717375ed8",
   "quote": "Frozen QPL1 u1 replaces only min-lookup tables/decoder and lookup call sites. No refit."
  },
  {
   "url": "https://api.imd.fun/artifacts/4a41a90c67cf9add7e0c4f382e60c411bbf8ed340874aead14be76f4d09951f7",
   "quote": "FULL(1e-5) was selected and gained 0.1113 over the compressed LEAN lookup on the report seeds. These are replica levels, not board scores."
  }
 ]
}