Back up verified OpenSysOne training snapshot and pinned source

.gitattributes

@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+source/results/20260917-playground/desktop.png filter=lfs diff=lfs merge=lfs -text
+source/results/20260917-playground/mobile.png filter=lfs diff=lfs merge=lfs -text

publications/20260917T022835Z-hf-backup/b05f7e3975a974c26ab667c6fdf91ec489a8cec0/manifest.json

@@ -0,0 +1,1329 @@
+{
+  "created_utc": "2026-09-17T03:38:54.741612+00:00",
+  "files": {
+    "snapshots/20260917T022835Z-hf-backup/SHA256SUMS": {
+      "sha256": "ba31bed97c788e6b514594575378327ad490657c3e51ff2c205af31e65be12e7",
+      "size": 4125
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/gx10-4b/best.pt": {
+      "sha256": "8956eb6c0cfbb02124aeefd99c3b418c55f55fdb9a64260350622d98dbba1aec",
+      "size": 66202635
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/gx10-4b/best_validation_selection.json": {
+      "sha256": "a458754f2b605558fecc8c6506349d9cdb1b7d6bd25fb4f701d481b1e0a778bb",
+      "size": 19698
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/gx10-4b/checkpoint.pt": {
+      "sha256": "5e869bb7fdb9abd84debc1040abf48f8f8c5a1437ebf0152f654ec7ed0754b5e",
+      "size": 198821629
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/gx10-4b/correctness_initial.json": {
+      "sha256": "5a3a0b5b7cd5b9ac2840fe9a6de37d60fc069f1958e315a8a0c973799a960198",
+      "size": 352
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/gx10-4b/data_filter.json": {
+      "sha256": "3a514e34a8a6f7e35776bafe5b19015d6c5cec0eaabea5558aa896471aba3b56",
+      "size": 2360
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/gx10-4b/manifest.json": {
+      "sha256": "8c81e2582258ec2c554b284e355e8059f317d91d24ec775e32dc86dc696c5845",
+      "size": 13020
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/gx10-4b/validation_step_002500_predictions.json": {
+      "sha256": "16ff7745cdbfa53b87ccf2445796cf6a0c28d148a00e6c3926c72c36e75d058e",
+      "size": 330419
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-a-4b/best.pt": {
+      "sha256": "b7923de78e41b88c353d27a13c7c66a0d5700aa578512a1ba3e5e34803486fb7",
+      "size": 66202635
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-a-4b/best_validation_selection.json": {
+      "sha256": "7a2dab3cda6e2b0d88cf4b9d7e5a82d3746b008e83e26eaba57316f92458bd5e",
+      "size": 19697
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-a-4b/checkpoint.pt": {
+      "sha256": "5a51c846bba8cb2cc83111fd4bca4f852f8b519ddbc84b702cf12ee6963ebb39",
+      "size": 198821629
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-a-4b/correctness_initial.json": {
+      "sha256": "cb2649db1735dbcbc90f83d9f95520a7dd553471b7d0c88316205207ad66acb8",
+      "size": 355
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-a-4b/data_filter.json": {
+      "sha256": "3a514e34a8a6f7e35776bafe5b19015d6c5cec0eaabea5558aa896471aba3b56",
+      "size": 2360
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-a-4b/manifest.json": {
+      "sha256": "839a92247fa8a861e853e46a18a97cebb9b4514b327562dc59066b8363fdf06a",
+      "size": 12984
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-a-4b/validation_step_002500_predictions.json": {
+      "sha256": "9d0676e4f318f0e419d4e73ce6b73afdc21a6252f751d8d88abcdd644b6ca979",
+      "size": 332103
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-2b/best.pt": {
+      "sha256": "be08efdc9121a64547b5c9331d5908eeb0978c4fbc962885fd34f9758ec89d11",
+      "size": 67383903
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-2b/best_validation_selection.json": {
+      "sha256": "030ee97e8df65bfd99781246c93bbab27c3e6218b92edee6cc40841923256df3",
+      "size": 19700
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-2b/checkpoint.pt": {
+      "sha256": "ab70dcb62432039d93ec71e200157250dc59c563404237e4dfad3662fd30f44f",
+      "size": 202309085
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-2b/correctness_final.json": {
+      "sha256": "dc12c9800ce68562e1fffe9d238170e7bd8a5f4400509dbb3a7207c2383c7522",
+      "size": 391
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-2b/correctness_initial.json": {
+      "sha256": "39d23512e99743437b88b5b10110e501853976b5292581481b5d2a5b7f2c5cdd",
+      "size": 391
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-2b/data_filter.json": {
+      "sha256": "e60da06955fe9e8a70a3de1ddaeb261e53af8f3a5e1d13ee62932d3773a1be24",
+      "size": 1572
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-2b/manifest.json": {
+      "sha256": "e6f39dc5c797a0c1f5aa6ca176509d4cb1ef176a98af0bb6ee0946a5acbbc09e",
+      "size": 10911
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-2b/validation_step_002000_predictions.json": {
+      "sha256": "980c9df18461f5de79e54701bfee1b5fbff19dd3b85125e64f619a88c60f80e4",
+      "size": 333354
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-4b-refinement/best.pt": {
+      "sha256": "a2b38db7e46440cdef7e5dee4dea5cc1f8310105abc1172c4c765900da4ce6a9",
+      "size": 66201995
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-4b-refinement/best_validation_selection.json": {
+      "sha256": "a458754f2b605558fecc8c6506349d9cdb1b7d6bd25fb4f701d481b1e0a778bb",
+      "size": 19698
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-4b-refinement/checkpoint.pt": {
+      "sha256": "ece94b096b7fa37eec9e180a5f652da86a9f4f42d49f7f124850f046915237a8",
+      "size": 198819453
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-4b-refinement/correctness_final.json": {
+      "sha256": "f9ad8b0308a09577abc390412f214cc6289b1108ced3907570c6d4cf58539158",
+      "size": 354
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-4b-refinement/correctness_initial.json": {
+      "sha256": "fd1b0d769f7a3f1cddacf00c19b0e3cec76c97c1d1e83a434ec2950bf178c677",
+      "size": 355
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-4b-refinement/data_filter.json": {
+      "sha256": "3a514e34a8a6f7e35776bafe5b19015d6c5cec0eaabea5558aa896471aba3b56",
+      "size": 2360
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-4b-refinement/initial_validation_predictions.json": {
+      "sha256": "16ff7745cdbfa53b87ccf2445796cf6a0c28d148a00e6c3926c72c36e75d058e",
+      "size": 330419
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-4b-refinement/initial_validation_selection.json": {
+      "sha256": "a458754f2b605558fecc8c6506349d9cdb1b7d6bd25fb4f701d481b1e0a778bb",
+      "size": 19698
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-4b-refinement/manifest.json": {
+      "sha256": "e1ac6aaf86db2fff6c007a1018503877976ba04de7415125594142158b6f11a5",
+      "size": 12394
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-4b-refinement/summary.json": {
+      "sha256": "929d4851d5b829a4cd69d337ab858f4f78480d6a654968d83141583c846fcec4",
+      "size": 11392
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-4b-refinement/training.jsonl": {
+      "sha256": "6865669d51de01930927a807d94d34b78499004fbf6dfd78cba55ab26554ecb7",
+      "size": 2114
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-4b-refinement/validation.jsonl": {
+      "sha256": "704b478c575080d30e1e555303b96b9307212d91c071673137f602a5b1db9fb6",
+      "size": 21145
+    },
+    "snapshots/20260917T022835Z-hf-backup/artifacts/spark-b-4b-refinement/validation_step_000008_predictions.json": {
+      "sha256": "b2a477b919ef21569299b4adf3b8c6d78806659e190dee622fbf393b16d79905",
+      "size": 330420
+    },
+    "snapshots/20260917T022835Z-hf-backup/backup-manifest.json": {
+      "sha256": "21c9df884c8362a97257997c6d5a62cf0d14bfc88fd46b7c782178cfce3c2ea8",
+      "size": 41326
+    },
+    "snapshots/20260917T022835Z-hf-backup/data/public-decisions-v1-20260916-manifest.json": {
+      "sha256": "adf5a8ca2bab60cf7429a82b1b7a2d1ae7d3cc812de04dd195a478dfd7c6f628",
+      "size": 5266
+    },
+    "source/.gitignore": {
+      "sha256": "c08a0eb968c43d53c1e54ef3f10e21ab4671c595cf25a2caa240b847acd06ddd",
+      "size": 73
+    },
+    "source/AGENTS.md": {
+      "sha256": "d27c6c6c5720db9c57efb1ee50a756c3d0486351fa2dcb33fd06a64fe7df7ac7",
+      "size": 1382
+    },
+    "source/FLEET_RUN.md": {
+      "sha256": "778339b8abd0dc1fa40d8c8b999a1cf39719a78cce8069bbf5b3f6bfa717b5be",
+      "size": 13851
+    },
+    "source/FLEET_SCOUT.md": {
+      "sha256": "789b020758a2713c74003f7d8d7af2d099c348d786c58c54136738a7ddcf7285",
+      "size": 8383
+    },
+    "source/HANDOVER.md": {
+      "sha256": "c46002891db5c557e28846a4f0adb89a531987c920da89a101493065b7bdfc24",
+      "size": 19142
+    },
+    "source/HF_MODEL_CARD.md": {
+      "sha256": "217a55e4c411423f93e12286c2b424d2ad886c3c5ae7f3092248bd08503dda5a",
+      "size": 6161
+    },
+    "source/HUGGINGFACE.md": {
+      "sha256": "521ae9303dfc69b65ba570811814c758434565e0b8057da97a135dcb5d7aab05",
+      "size": 5610
+    },
+    "source/JEV_HARNESS.md": {
+      "sha256": "c0f4a0d9055689b6a8be6fb66aadb43ed448609002618dd4d7e3857808c5e88b",
+      "size": 4683
+    },
+    "source/NEXT_STEPS.md": {
+      "sha256": "8fc19584153ca56b3e86ca77af7dd01e6d5a8d05a998973359dc003e5ff56a4f",
+      "size": 12181
+    },
+    "source/PLAN.md": {
+      "sha256": "9ec4049ed36158a8d176ac8f18b7ff16908ba5ab7f457188f7ccf2a4dfb5b820",
+      "size": 15930
+    },
+    "source/PLAYGROUND.md": {
+      "sha256": "506486e66e33e74925cd78092f02c0f50a1d163b2bcf5858bc52a2c4478352e5",
+      "size": 4790
+    },
+    "source/README.md": {
+      "sha256": "f5e5afde7c7073bd5f6e482e9f8eb61f8cd76268293a13626d6bc812a6bdc812",
+      "size": 3740
+    },
+    "source/RESEARCH_BRIEF.md": {
+      "sha256": "9ed40f7da30fbc0d1492d30e6bd5a2f315167f077293cbc3f7345de3a20ef164",
+      "size": 24894
+    },
+    "source/RESEARCH_NOTES.md": {
+      "sha256": "e11d9f1e09481e741cd333d7fe84e566bd468db40af77d92f5c361f4c428306c",
+      "size": 13713
+    },
+    "source/RESULTS.md": {
+      "sha256": "18d688c4e25f5e7008a9c6ae89da1e68370ec8b1d9c5b181f56908e4db22281d",
+      "size": 30956
+    },
+    "source/decision_model.py": {
+      "sha256": "a3d8aeb02a1ac765c6cc30ff175acad0664560f01ab5403e22cade924d17371e",
+      "size": 9258
+    },
+    "source/examples/jev_request.json": {
+      "sha256": "74f07501aa665284ab0611b6a3ed1fdc046821be10dddaa0a503efc52d7d7eb5",
+      "size": 703
+    },
+    "source/experiment.py": {
+      "sha256": "c036cbe9fbf6f71c1bb141dcfbecfb06a2f4bd697b070d2274b1931480e92031",
+      "size": 42981
+    },
+    "source/jev_harness.py": {
+      "sha256": "4d4e979cb7ae352bcdacaaa6d64045e6b5e550b1a9721d4bad545045bee6c67f",
+      "size": 16898
+    },
+    "source/playground.py": {
+      "sha256": "0eed2ab323dc73c7970b6d11fe41ddcb25e03f376af7e27891fd7e6c6a1e6e12",
+      "size": 18522
+    },
+    "source/results/20260916-fleet-setup/Qwen3-4B-Instruct-2507-cdbee75f-files.json": {
+      "sha256": "3ceddf4e5228246ddb9a6ad381fb31793a61b050a8fb10934a345225e6068fe6",
+      "size": 1971
+    },
+    "source/results/20260916-fleet-setup/Qwen3.5-2B-15852e8c-files.json": {
+      "sha256": "a78fb2654833d89f5b274b562a4595f2dff001065db2cf64f83bcd1a206c7576",
+      "size": 1950
+    },
+    "source/results/20260916-fleet-setup/cpu-tests.json": {
+      "sha256": "2110698e40353ee830b3f7bf792c5f2c8ea0aff54c8fa7d9b6b49fde99540b71",
+      "size": 328
+    },
+    "source/results/20260916-fleet-setup/crossfit-tests.json": {
+      "sha256": "a804cfb5eaf7a963085cb399da2f0789f8a50c56bd6c38f44b7906c4fffe2d3f",
+      "size": 532
+    },
+    "source/results/20260916-fleet-setup/evidence-index.json": {
+      "sha256": "db4d25099b45af53cffa54551cb0c56f82b1a0228915455873183f31a228148e",
+      "size": 12480
+    },
+    "source/results/20260916-fleet-setup/fleet-current-status.json": {
+      "sha256": "06b07b3ed301bdfaacfe4ded988033e525e4ea7580b0f6bdbe6bdaceb4a1bab9",
+      "size": 11377
+    },
+    "source/results/20260916-fleet-setup/fleet-launch-verification.json": {
+      "sha256": "9f6ee33c3004645c851033f87b8fb748478d37a2b59dff75b35fff9e1f45a413",
+      "size": 553
+    },
+    "source/results/20260916-fleet-setup/fleet-plan.json": {
+      "sha256": "bff9ed43f7fd7710ec3915144424daf1f0ffb1160ae86ee3a654cbba7974cd4b",
+      "size": 1795
+    },
+    "source/results/20260916-fleet-setup/fleet-state.json": {
+      "sha256": "a84107d6c7ee5c96995cd3fd314d5ce217c1ef29f6296f8771b3bdfdaa6fc018",
+      "size": 664
+    },
+    "source/results/20260916-fleet-setup/gx10-current-best_validation_predictions.json": {
+      "sha256": "9f1f8ccbc26a29ba0a0615ff1cbf278776c01c003b63176bae41df5a5bde92e6",
+      "size": 332042
+    },
+    "source/results/20260916-fleet-setup/gx10-current-best_validation_selection.json": {
+      "sha256": "9cbc9717355d4d8e17d1d072cae9c243c37e2497e691ca0685fe5b6194c3d74b",
+      "size": 19696
+    },
+    "source/results/20260916-fleet-setup/gx10-current-correctness_initial.json": {
+      "sha256": "5a3a0b5b7cd5b9ac2840fe9a6de37d60fc069f1958e315a8a0c973799a960198",
+      "size": 352
+    },
+    "source/results/20260916-fleet-setup/gx10-current-manifest.json": {
+      "sha256": "8c81e2582258ec2c554b284e355e8059f317d91d24ec775e32dc86dc696c5845",
+      "size": 13020
+    },
+    "source/results/20260916-fleet-setup/gx10-current-startup-verification.json": {
+      "sha256": "875e7c410ad0c98328737c2f34b5da1d9dedf9d5ba44cfed0b9bdd43226f73af",
+      "size": 24964
+    },
+    "source/results/20260916-fleet-setup/gx10-current-validation_step_000178_predictions.json": {
+      "sha256": "9f1f8ccbc26a29ba0a0615ff1cbf278776c01c003b63176bae41df5a5bde92e6",
+      "size": 332042
+    },
+    "source/results/20260916-fleet-setup/gx10-python-stack-proof.json": {
+      "sha256": "9ee3222204613414ecbcba0a2faecf740c03e1d771fbdcc0af5ff5d98a5bb62a",
+      "size": 364
+    },
+    "source/results/20260916-fleet-setup/gx10-resume-state-verification.json": {
+      "sha256": "5ce557a67fd757854667e36473754c6b14928615c9a6564bc036fe3bd0e1465c",
+      "size": 635
+    },
+    "source/results/20260916-fleet-setup/gx10-transition-checkpoint.json": {
+      "sha256": "f14a3b2da291655c3296e1a5fad439c83de1cf7e143318e74781b4fb1767f6a6",
+      "size": 581
+    },
+    "source/results/20260916-fleet-setup/selection-diagnostic.json": {
+      "sha256": "783190fe970f0930b449632065b2689d5b2e91a6d7a31415cf4710c9907ba55f",
+      "size": 4482
+    },
+    "source/results/20260916-fleet-setup/selection_migration.json": {
+      "sha256": "37bc6f9606d3e204bcdf45e0cd50bea2e5713f8d639c98fbd39f92c665c527bc",
+      "size": 1389
+    },
+    "source/results/20260916-fleet-setup/spark-a-best-validation-selection.json": {
+      "sha256": "e6cd9400edd4d510fe0342f726aead056c075c0829a767400e1ffd335dc86b39",
+      "size": 19697
+    },
+    "source/results/20260916-fleet-setup/spark-a-campaign-plan.json": {
+      "sha256": "f04bbd72d5a7cc59980714f1ef674ec7c9af309d8a956a7e5a03b766c647af74",
+      "size": 777
+    },
+    "source/results/20260916-fleet-setup/spark-a-campaign-state-snapshot.json": {
+      "sha256": "417d9f1552cfa82ddc8ad49224b986ef5d58f89d942a73b338a6aa6646cb4ab9",
+      "size": 1421
+    },
+    "source/results/20260916-fleet-setup/spark-a-campaign-updates-verified.json": {
+      "sha256": "52eae7f877102c4c46a9cf3275a354784f0c95f13c9772cc578db12c101a7698",
+      "size": 4674
+    },
+    "source/results/20260916-fleet-setup/spark-a-candidate-launch.json": {
+      "sha256": "e35162501a6d207f7467b44cc27b3d3fb1133db4f212a0b84942271909bdda91",
+      "size": 815
+    },
+    "source/results/20260916-fleet-setup/spark-a-inherited-validation-selection.json": {
+      "sha256": "2659b2b3fe8b7c8d9e614b30c25a99ad7100cd007892fba9b4ea6629fde0d887",
+      "size": 19687
+    },
+    "source/results/20260916-fleet-setup/spark-a-initial-validation-selection.json": {
+      "sha256": "e6cd9400edd4d510fe0342f726aead056c075c0829a767400e1ffd335dc86b39",
+      "size": 19697
+    },
+    "source/results/20260916-fleet-setup/spark-a-inputs-verified.json": {
+      "sha256": "7e6e8642a752993aa7181dcfb96e016ca4dbdf454109ff99d7174efd5a3c6f5b",
+      "size": 2202
+    },
+    "source/results/20260916-fleet-setup/spark-a-launcher.json": {
+      "sha256": "422197489b6de521d73d746a8bf21cfabc9ac82da82dfa892cfd68ce3526fd34",
+      "size": 140
+    },
+    "source/results/20260916-fleet-setup/spark-a-model-copy-verified.json": {
+      "sha256": "ffed13b2c1945c091bf2ea586556f94ba5e1ff67e7aeb92e1cbd49c488ae2ef9",
+      "size": 2188
+    },
+    "source/results/20260916-fleet-setup/spark-a-pilot-updates-verified.json": {
+      "sha256": "0ba2ad5967dbc5fc2e7a01a4bbf7a4ec6f1a7fe50bb815d9cefc63f314d1b154",
+      "size": 2897
+    },
+    "source/results/20260916-fleet-setup/spark-a-pilot/best_validation_predictions.json": {
+      "sha256": "e671e1508185765552b0f933ba03f356be62143c531d8ef534457d34b1645c9b",
+      "size": 333136
+    },
+    "source/results/20260916-fleet-setup/spark-a-pilot/correctness_final.json": {
+      "sha256": "cb2649db1735dbcbc90f83d9f95520a7dd553471b7d0c88316205207ad66acb8",
+      "size": 355
+    },
+    "source/results/20260916-fleet-setup/spark-a-pilot/correctness_initial.json": {
+      "sha256": "5d6334a0c66ed567fa9709df4ea5b6040c88bb5258e864332e88be2bb9586634",
+      "size": 355
+    },
+    "source/results/20260916-fleet-setup/spark-a-pilot/data_filter.json": {
+      "sha256": "3a514e34a8a6f7e35776bafe5b19015d6c5cec0eaabea5558aa896471aba3b56",
+      "size": 2360
+    },
+    "source/results/20260916-fleet-setup/spark-a-pilot/initial_validation_predictions.json": {
+      "sha256": "e671e1508185765552b0f933ba03f356be62143c531d8ef534457d34b1645c9b",
+      "size": 333136
+    },
+    "source/results/20260916-fleet-setup/spark-a-pilot/manifest.json": {
+      "sha256": "18672b6c49227a08ba1303a4958a9fa08ae78b5fe59e30c99902e5d61ccaddf0",
+      "size": 12238
+    },
+    "source/results/20260916-fleet-setup/spark-a-pilot/summary.json": {
+      "sha256": "9bda8ed53ed42f7a8b4dea08acbb6c4c58635bf64dafa925b1a3575e9a4f89b4",
+      "size": 11499
+    },
+    "source/results/20260916-fleet-setup/spark-a-pilot/training.jsonl": {
+      "sha256": "fe3be4437a159b98ddc84fb2da04f42de44aa10eebd8e19029d92f4a9585bc8f",
+      "size": 2082
+    },
+    "source/results/20260916-fleet-setup/spark-a-pilot/validation.jsonl": {
+      "sha256": "ef002be6c6dd5a27fa35f555d0a3fb3544f885ed806e026edeb2491fbb9ccd84",
+      "size": 6191
+    },
+    "source/results/20260916-fleet-setup/spark-a-pilot/validation_step_000008_predictions.json": {
+      "sha256": "d520b46127014161384c46d856fcfa167d4bb667a369235678f2e96ab0e54029",
+      "size": 333027
+    },
+    "source/results/20260916-fleet-setup/spark-a-python-stack-verified.json": {
+      "sha256": "8391cccb0d72f8a9af7e631075435a83ba3582c8bdd30b6ead0a3e50d90c46d3",
+      "size": 1965
+    },
+    "source/results/20260916-fleet-setup/spark-a-resume-state-verified.json": {
+      "sha256": "18ad5ae7e27f1d7c20b4de4cc0537fd1d77cf665f00ec91c87df9872c328d3c6",
+      "size": 1008
+    },
+    "source/results/20260916-fleet-setup/spark-a-resume-verified.json": {
+      "sha256": "7bbd348d2aed3478ec8967913068ce50c85acc6853d2dd33c56e1772c26369f0",
+      "size": 625
+    },
+    "source/results/20260916-fleet-setup/spark-a-service-stop.json": {
+      "sha256": "b2091baa22a19691fd9e5f4ce17c4bf61bcc07f2bdecd67b42fcd8e22b9af907",
+      "size": 1923
+    },
+    "source/results/20260916-fleet-setup/spark-a-setup-status.json": {
+      "sha256": "9ebe7f313f592080899534e55fda097aece13e559bf6fde67a0489f191cbe50d",
+      "size": 102
+    },
+    "source/results/20260916-fleet-setup/spark-a-training-correctness-initial.json": {
+      "sha256": "cb2649db1735dbcbc90f83d9f95520a7dd553471b7d0c88316205207ad66acb8",
+      "size": 355
+    },
+    "source/results/20260916-fleet-setup/spark-a-training-manifest.json": {
+      "sha256": "839a92247fa8a861e853e46a18a97cebb9b4514b327562dc59066b8363fdf06a",
+      "size": 12984
+    },
+    "source/results/20260916-fleet-setup/spark-a-verification/data_filter.json": {
+      "sha256": "3a514e34a8a6f7e35776bafe5b19015d6c5cec0eaabea5558aa896471aba3b56",
+      "size": 2360
+    },
+    "source/results/20260916-fleet-setup/spark-a-verification/http_255_choices_response.json": {
+      "sha256": "36ce43c4fc4885a73daa7f57c27143f311d0a075621e8d2f1121a4cf64b815f4",
+      "size": 12050
+    },
+    "source/results/20260916-fleet-setup/spark-a-verification/http_long_context_response.json": {
+      "sha256": "762a8383eaec4cc0ca22738d8cc9abde026781c37468ec2b5bf3058fd0e487d5",
+      "size": 216
+    },
+    "source/results/20260916-fleet-setup/spark-a-verification/http_response.json": {
+      "sha256": "f43f195c9ac3b62714e4dc11f99b71ac7e56aec89f979a481298b8039bd7a24f",
+      "size": 859
+    },
+    "source/results/20260916-fleet-setup/spark-a-verification/manifest.json": {
+      "sha256": "6a598e188f2d829ac3b3f7762fd67047691ba2246bddf959e3a073b8640986f7",
+      "size": 264
+    },
+    "source/results/20260916-fleet-setup/spark-a-verification/reload_predictions.json": {
+      "sha256": "f56623fbe5d0c26e6433bea78c77ecd61f4459f1ae3b0b62b808dae4a836b2fd",
+      "size": 10480
+    },
+    "source/results/20260916-fleet-setup/spark-a-verification/stress.json": {
+      "sha256": "5471c772bf1f56ad0c2df20c329eccf77aa0161eeb5953295c4959c9e298cf06",
+      "size": 1169
+    },
+    "source/results/20260916-fleet-setup/spark-a-verification/verification.json": {
+      "sha256": "3835333780af23e0c3398aa22b4b0da207a399e26ffa868978a457c4370a7f02",
+      "size": 2300
+    },
+    "source/results/20260916-fleet-setup/spark-a-warmstart-verified.json": {
+      "sha256": "516d935684a1a747ea6d74389568507a2fe49743a4d5b11dca89524f5992eaaf",
+      "size": 634
+    },
+    "source/results/20260916-fleet-setup/spark-b-best-validation-selection.json": {
+      "sha256": "9d16f6144a3ff81ffbd1a6cf78078d4bd01fa683f8bbfd4028627591686958f6",
+      "size": 19690
+    },
+    "source/results/20260916-fleet-setup/spark-b-correctness-initial.json": {
+      "sha256": "39d23512e99743437b88b5b10110e501853976b5292581481b5d2a5b7f2c5cdd",
+      "size": 391
+    },
+    "source/results/20260916-fleet-setup/spark-b-current-status.json": {
+      "sha256": "00af691a392d04dc2b8359a4232f836ec8f96501153025acc91ff370bcc9f357",
+      "size": 2861
+    },
+    "source/results/20260916-fleet-setup/spark-b-data-filter.json": {
+      "sha256": "e60da06955fe9e8a70a3de1ddaeb261e53af8f3a5e1d13ee62932d3773a1be24",
+      "size": 1572
+    },
+    "source/results/20260916-fleet-setup/spark-b-evidence-index.json": {
+      "sha256": "8b6e4a1172667c614749a3135d060183eefbf701a4f0c011a57b452d0d9b4490",
+      "size": 6279
+    },
+    "source/results/20260916-fleet-setup/spark-b-http-255-choices-response.json": {
+      "sha256": "fe9c590782a9584a7ccca84efcce728918e6a0d53edb11a86de4ca06a9d0ec9c",
+      "size": 11977
+    },
+    "source/results/20260916-fleet-setup/spark-b-http-long-context-response.json": {
+      "sha256": "77537e1b07b9746e7c75c503de97d570bbedfdc10c4e3be602d9aca7dfa6fd98",
+      "size": 202
+    },
+    "source/results/20260916-fleet-setup/spark-b-http-response.json": {
+      "sha256": "2baaeafd62463d40bd7318f0a24bfa423897f1139de2c60ccd16d072fd4955c6",
+      "size": 844
+    },
+    "source/results/20260916-fleet-setup/spark-b-inherited-validation-selection.json": {
+      "sha256": "9d16f6144a3ff81ffbd1a6cf78078d4bd01fa683f8bbfd4028627591686958f6",
+      "size": 19690
+    },
+    "source/results/20260916-fleet-setup/spark-b-initial-validation-selection.json": {
+      "sha256": "9d16f6144a3ff81ffbd1a6cf78078d4bd01fa683f8bbfd4028627591686958f6",
+      "size": 19690
+    },
+    "source/results/20260916-fleet-setup/spark-b-model-copy-verified.json": {
+      "sha256": "d6c5d97275d80f93e0a0713f3801eed6b173046d9d5c95989fa91504ad5c5b3f",
+      "size": 2155
+    },
+    "source/results/20260916-fleet-setup/spark-b-plan.json": {
+      "sha256": "509505b3fcb2af1d50b16bfdecd5833a7691da217700923d080aced6d679016f",
+      "size": 777
+    },
+    "source/results/20260916-fleet-setup/spark-b-python-stack-verified.json": {
+      "sha256": "c1eecf2c6d3c1a664180abdc99161a2e0372af5b0dbfab4fc354ebb49a5abfb6",
+      "size": 1965
+    },
+    "source/results/20260916-fleet-setup/spark-b-reload-predictions.json": {
+      "sha256": "4cc860ca3e6a65a904d17609d74102e39283480fd4f1e5f69cee4d8c7dc135bb",
+      "size": 10455
+    },
+    "source/results/20260916-fleet-setup/spark-b-service-stop.json": {
+      "sha256": "8706294be28526fc6fb1ece61964a61085a32636e5cd399387a2bd6e3dd7012f",
+      "size": 1201
+    },
+    "source/results/20260916-fleet-setup/spark-b-setup-launch.json": {
+      "sha256": "a065853c9ad9d6f63475fcc4996b289c7bb914bf50129e06b38d820145499347",
+      "size": 915
+    },
+    "source/results/20260916-fleet-setup/spark-b-setup-stages.log": {
+      "sha256": "13216b31b4287dc33421e7a8264133b5615a15e463b45fffa740eacfd4dd843d",
+      "size": 260
+    },
+    "source/results/20260916-fleet-setup/spark-b-setup-status.json": {
+      "sha256": "3f11175e52779f7ffb1e1848f6a57e859d1a318b10c8424d2695464d060a1a82",
+      "size": 102
+    },
+    "source/results/20260916-fleet-setup/spark-b-startup-verification.json": {
+      "sha256": "9eeef85649a5765a9a465fd082d5ef71687afbe1c7c910b6b1395507b3faeb94",
+      "size": 9581
+    },
+    "source/results/20260916-fleet-setup/spark-b-stress.json": {
+      "sha256": "617709fb02a0098b6ccbfce17d5817cbacf680178cd0a2d50e6c02557b2160d8",
+      "size": 1162
+    },
+    "source/results/20260916-fleet-setup/spark-b-training-manifest.json": {
+      "sha256": "e6f39dc5c797a0c1f5aa6ca176509d4cb1ef176a98af0bb6ee0946a5acbbc09e",
+      "size": 10911
+    },
+    "source/results/20260916-fleet-setup/spark-b-verification-manifest.json": {
+      "sha256": "891e5a059daa63397bcd403bf4f2e4297c5aba983272d25b316ac6a68c3b9457",
+      "size": 264
+    },
+    "source/results/20260916-fleet-setup/spark-b-verification.json": {
+      "sha256": "0ee15bcf7d01d312b91ec935fefa9944338847cf8a1f67cabba79e1a141c0af3",
+      "size": 2279
+    },
+    "source/results/20260916T154714Z/base_token_yes_minus_no_predictions.json": {
+      "sha256": "5c6dc5ab1c21d68cff293fbf73cdea66447d9a7ad2c3aaaac976608c7a696ecd",
+      "size": 31072
+    },
+    "source/results/20260916T154714Z/calibrated_predictions.json": {
+      "sha256": "cfd85bd7ae72ac090f326a8fe534b242230dad9b8772ddb50d3bddda50e25e70",
+      "size": 33872
+    },
+    "source/results/20260916T154714Z/calibration.jsonl": {
+      "sha256": "09e51eeffa7cdc2faad78f3dc6fd7b994580aae7e0ac2ecf8f1ceafdd7f984a0",
+      "size": 16193
+    },
+    "source/results/20260916T154714Z/calibration_predictions.json": {
+      "sha256": "f0d1fb5ba5d9778ccc4db37830372ec2257ce76f6a082eb63b00ee6009f768be",
+      "size": 23027
+    },
+    "source/results/20260916T154714Z/exit_code": {
+      "sha256": "4355a46b19d348dc2f57c046f8ef63d4538ebb936000f3c9ee954a27460dd865",
+      "size": 2
+    },
+    "source/results/20260916T154714Z/initial_scalar_predictions.json": {
+      "sha256": "6aa8afae1afb92d4ebf273d80741f3b165e694a20280a1da2e9144425c2d8294",
+      "size": 28395
+    },
+    "source/results/20260916T154714Z/manifest.json": {
+      "sha256": "4b4a55380d16561b0703d2b6c1d94b05a7f4d129b193656c1003dc80548e6414",
+      "size": 3462
+    },
+    "source/results/20260916T154714Z/parity_diagnosis.json": {
+      "sha256": "060e44827c102ed8c4f5a317dea9822a94fa8d87d8d795fb4a596e9954947caa",
+      "size": 6482
+    },
+    "source/results/20260916T154714Z/run.log": {
+      "sha256": "1659885bb673aed6f5ad0249111c75edab218bb46aefa7e2d3c38f2bc381448c",
+      "size": 11626
+    },
+    "source/results/20260916T154714Z/test.jsonl": {
+      "sha256": "4c0c365566099ff0941b2da473d6d3cca445ef1174c931d9d93f9bb475679599",
+      "size": 22779
+    },
+    "source/results/20260916T154714Z/train.jsonl": {
+      "sha256": "5ff456f912cf22ddefd8e0e24baf3956f0f010cd4c605e186a36659b4b931ea0",
+      "size": 61325
+    },
+    "source/results/20260916T154714Z/trained_predictions.json": {
+      "sha256": "8ebd0aad15bd04b4399cb467c4bad2bca813345179f737fbe510031d05a2e8a8",
+      "size": 34043
+    },
+    "source/results/20260916T154714Z/training.json": {
+      "sha256": "e3b0a701faa2eb67a3bbb70f4662e3f71d1587a396443afa709ec5e548577723",
+      "size": 10673
+    },
+    "source/results/20260916T155124Z/base_token_yes_minus_no_predictions.json": {
+      "sha256": "14bdc2ffad677e3fd3b6c58fae74efe3a7530df29ac9e181c2379a0cd5085ce1",
+      "size": 33837
+    },
+    "source/results/20260916T155124Z/benchmark.json": {
+      "sha256": "01027c6338ff2e36e3b7808f47c4f279bc9a2e52f24efc0bf7fd5744b5b572d8",
+      "size": 7692
+    },
+    "source/results/20260916T155124Z/calibrated_predictions.json": {
+      "sha256": "807e1f999a9ee64fa5f93851d1ee29da2d0fd41ed23e44e46f6391fb1f822a2c",
+      "size": 33865
+    },
+    "source/results/20260916T155124Z/calibration.jsonl": {
+      "sha256": "09e51eeffa7cdc2faad78f3dc6fd7b994580aae7e0ac2ecf8f1ceafdd7f984a0",
+      "size": 16193
+    },
+    "source/results/20260916T155124Z/calibration_predictions.json": {
+      "sha256": "1f3a0277c66f27c5ff08028ce6a7ccd90b9ccd5f7105b74fb8a3ae69c532422d",
+      "size": 23026
+    },
+    "source/results/20260916T155124Z/correctness.json": {
+      "sha256": "566f735523c8d1e46e4b11534b436d11edf1a9b1aef90e380aa10024b2b19abe",
+      "size": 329
+    },
+    "source/results/20260916T155124Z/exit_code": {
+      "sha256": "9a271f2a916b0b6ee6cecb2426f0b3206ef074578be55d9bc94f6f3fe3ab86aa",
+      "size": 2
+    },
+    "source/results/20260916T155124Z/initial_scalar_predictions.json": {
+      "sha256": "6aa8afae1afb92d4ebf273d80741f3b165e694a20280a1da2e9144425c2d8294",
+      "size": 28395
+    },
+    "source/results/20260916T155124Z/manifest.json": {
+      "sha256": "51b417c25f4b6f169b214bd6e4662a59b59365fc7e048feaa555941a05208a61",
+      "size": 3609
+    },
+    "source/results/20260916T155124Z/metrics.json": {
+      "sha256": "f5ecee19932286363c1e95bdb91c977dc907c9dc2a0f2bee85cfbb9c01472738",
+      "size": 6152
+    },
+    "source/results/20260916T155124Z/run.log": {
+      "sha256": "e0589922a98bdf412f749004cae387f8f60ef9c6f43f983de2cf433f6d41b6bc",
+      "size": 17478
+    },
+    "source/results/20260916T155124Z/test.jsonl": {
+      "sha256": "4c0c365566099ff0941b2da473d6d3cca445ef1174c931d9d93f9bb475679599",
+      "size": 22779
+    },
+    "source/results/20260916T155124Z/train.jsonl": {
+      "sha256": "5ff456f912cf22ddefd8e0e24baf3956f0f010cd4c605e186a36659b4b931ea0",
+      "size": 61325
+    },
+    "source/results/20260916T155124Z/trained_predictions.json": {
+      "sha256": "a0a020c3ec7134e82eab3fe22fe2e533d0dfe2393bf2e2f21c26020c09e5ba2a",
+      "size": 34016
+    },
+    "source/results/20260916T155124Z/training.json": {
+      "sha256": "6586a74826b7c0c7423a673cb7d9f6ce753dd097513802da18f975d769c9620a",
+      "size": 10643
+    },
+    "source/results/20260916T155314Z/benchmark.json": {
+      "sha256": "74fe266abe96c407487ccbe71ad4d30b78990b8c3382b68f64063ce76f85ee28",
+      "size": 7688
+    },
+    "source/results/20260916T155314Z/calibrated_predictions.json": {
+      "sha256": "aa4ffbf0fd363d2ba5a7f6d4ccff355dc636e88f8ef1b723eb2b0426e73aac46",
+      "size": 33876
+    },
+    "source/results/20260916T155314Z/calibration.jsonl": {
+      "sha256": "09e51eeffa7cdc2faad78f3dc6fd7b994580aae7e0ac2ecf8f1ceafdd7f984a0",
+      "size": 16193
+    },
+    "source/results/20260916T155314Z/calibration_predictions.json": {
+      "sha256": "eb2b47919f765cef2eb6d38ea1aff23e17ec2691f3b8052ee205b126130d7905",
+      "size": 22993
+    },
+    "source/results/20260916T155314Z/correctness.json": {
+      "sha256": "faf4077e01a6b0fede1988c98a586d71fc81c454b621a8159f9206d097aacce3",
+      "size": 335
+    },
+    "source/results/20260916T155314Z/exit_code": {
+      "sha256": "9a271f2a916b0b6ee6cecb2426f0b3206ef074578be55d9bc94f6f3fe3ab86aa",
+      "size": 2
+    },
+    "source/results/20260916T155314Z/manifest.json": {
+      "sha256": "51d594b45c7d67db6ea0be5a61d9fdf1d3d8b7e2b6805630e0a5bb255b734761",
+      "size": 3725
+    },
+    "source/results/20260916T155314Z/metrics.json": {
+      "sha256": "f0d51ea2136aff729163daccd56435636b751633a826f1e834cd4b332fb8e880",
+      "size": 4966
+    },
+    "source/results/20260916T155314Z/resume_verification.json": {
+      "sha256": "e9e80d9357d17c01f650f9b42cd358ed7193588736c98d91aaed21bc9ed30122",
+      "size": 773
+    },
+    "source/results/20260916T155314Z/resumed_initial_predictions.json": {
+      "sha256": "a0a020c3ec7134e82eab3fe22fe2e533d0dfe2393bf2e2f21c26020c09e5ba2a",
+      "size": 34016
+    },
+    "source/results/20260916T155314Z/run.log": {
+      "sha256": "05e8ef6eb53ff8fb01fb2571054501a4fd70b8d301e1dd35c81b248c02eb5769",
+      "size": 6914
+    },
+    "source/results/20260916T155314Z/test.jsonl": {
+      "sha256": "4c0c365566099ff0941b2da473d6d3cca445ef1174c931d9d93f9bb475679599",
+      "size": 22779
+    },
+    "source/results/20260916T155314Z/train.jsonl": {
+      "sha256": "5ff456f912cf22ddefd8e0e24baf3956f0f010cd4c605e186a36659b4b931ea0",
+      "size": 61325
+    },
+    "source/results/20260916T155314Z/trained_predictions.json": {
+      "sha256": "bdd004b8bb553cd446f159491f2be9c9e5e3f8f9b92061454ff20fb9850fcc92",
+      "size": 34004
+    },
+    "source/results/20260916T155314Z/training.json": {
+      "sha256": "6d56c059831aff521d59894702d7a66574d9e7ba8f965f745fa29e7531c5eb23",
+      "size": 180
+    },
+    "source/results/20260916T161253Z-precision/exit_code": {
+      "sha256": "9a271f2a916b0b6ee6cecb2426f0b3206ef074578be55d9bc94f6f3fe3ab86aa",
+      "size": 2
+    },
+    "source/results/20260916T161253Z-precision/manifest.json": {
+      "sha256": "7626809df6622a05507b11e65444e1d1b28e9a03ea81838169977991aabc6de5",
+      "size": 1169
+    },
+    "source/results/20260916T161253Z-precision/precision.json": {
+      "sha256": "5151de2f07119b2c6de007060899883f0cdb36ff728982df2b95de092d0097d8",
+      "size": 110342
+    },
+    "source/results/20260916T161253Z-precision/run.log": {
+      "sha256": "215e898ab1cf849c98ef56263e8d22a570d79263af2a39af925863adc394591f",
+      "size": 1655
+    },
+    "source/results/20260916T161355Z-precision/exit_code": {
+      "sha256": "9a271f2a916b0b6ee6cecb2426f0b3206ef074578be55d9bc94f6f3fe3ab86aa",
+      "size": 2
+    },
+    "source/results/20260916T161355Z-precision/manifest.json": {
+      "sha256": "93d8421244498c398f437adf4e6af0b4eb6b9b1a6f4d447564d1442faba56eae",
+      "size": 1103
+    },
+    "source/results/20260916T161355Z-precision/precision.json": {
+      "sha256": "d53276f38ba72bb01219d5e70183476cded0cd1b4bddb4cfb1ddddb9ec212000",
+      "size": 153559
+    },
+    "source/results/20260916T161355Z-precision/run.log": {
+      "sha256": "6696cda7c457d6c5c178d3adc4301b254dfe406b22dcfa8f0975c8b055755097",
+      "size": 2359
+    },
+    "source/results/20260916T182352Z-train/correctness_final.json": {
+      "sha256": "55e085c42e416d00fd4cdad9fb72d121f780fb9005226549ce7eaeddc9eea051",
+      "size": 408
+    },
+    "source/results/20260916T182352Z-train/correctness_initial.json": {
+      "sha256": "7414b27fe26d2d52ea023ba5cc4d8bc9a46d9bf60a8d54358ad2ad803c03c26a",
+      "size": 479
+    },
+    "source/results/20260916T182352Z-train/data_filter.json": {
+      "sha256": "e60da06955fe9e8a70a3de1ddaeb261e53af8f3a5e1d13ee62932d3773a1be24",
+      "size": 1572
+    },
+    "source/results/20260916T182352Z-train/initial_validation_predictions.json": {
+      "sha256": "4ee92802d7a8d642c8801dfd96f5bb18df80f745632fc1c8d99f9da9fe731c5b",
+      "size": 332151
+    },
+    "source/results/20260916T182352Z-train/manifest.json": {
+      "sha256": "e1efa2f4d350c678ba726e88dc85d8e7f3c8dc4e698742897892e578e5d94c58",
+      "size": 9460
+    },
+    "source/results/20260916T182352Z-train/resume_verification.json": {
+      "sha256": "501cb2750d0d5ba914dd066bd1312b39ec1e476ab8c074ac1b2758680f8a1928",
+      "size": 986
+    },
+    "source/results/20260916T182352Z-train/summary.json": {
+      "sha256": "05cf399210326916c92e404ed815ca8644bd3db09fa027cbfe4686fac47a902f",
+      "size": 11569
+    },
+    "source/results/20260916T182352Z-train/training.jsonl": {
+      "sha256": "231253f8237df8a40c95e6550fc36975b12b3b657e80da72c98a7712d365e48d",
+      "size": 10396
+    },
+    "source/results/20260916T182352Z-train/validation.jsonl": {
+      "sha256": "fdc0ad4f5d854f2ea33e45e644117e16d831d01576b9f011ccbca75745bc6c9b",
+      "size": 6202
+    },
+    "source/results/20260916T182352Z-train/validation_step_000040_predictions.json": {
+      "sha256": "d44337f6e32bd8f1e9db01f533ce031dafec5a6162fe43d19dde40e0a3c39c75",
+      "size": 332208
+    },
+    "source/results/20260916T183240Z-train/correctness_final.json": {
+      "sha256": "f941e3b1c39b5636b59ac513a9a0673ef683d708c52e505cb95b8271163efaf8",
+      "size": 391
+    },
+    "source/results/20260916T183240Z-train/correctness_initial.json": {
+      "sha256": "39d23512e99743437b88b5b10110e501853976b5292581481b5d2a5b7f2c5cdd",
+      "size": 391
+    },
+    "source/results/20260916T183240Z-train/manifest.json": {
+      "sha256": "ea6f9690fed03a43d2698161c3aae1685aca289cffd7f430e4ade5c32d16bb2d",
+      "size": 9818
+    },
+    "source/results/20260916T183240Z-train/summary.json": {
+      "sha256": "873e71eda53e3d000d805faa86d5a436611e97d74f331f89b9ed08d055f116a4",
+      "size": 11522
+    },
+    "source/results/20260916T183240Z-train/training.jsonl": {
+      "sha256": "7ea9596dd9503513d55715f4d0dee5f3aa8e22bd291288c4fde32511f8691677",
+      "size": 261
+    },
+    "source/results/20260916T183240Z-train/validation.jsonl": {
+      "sha256": "5f3e0dc2ed0e96ebff2673c637ff167c9f7d14691e1b98770f28b55467cc818a",
+      "size": 6190
+    },
+    "source/results/20260916T183751Z-verify2b/data_filter.json": {
+      "sha256": "e60da06955fe9e8a70a3de1ddaeb261e53af8f3a5e1d13ee62932d3773a1be24",
+      "size": 1572
+    },
+    "source/results/20260916T183751Z-verify2b/http_response.json": {
+      "sha256": "b8715eaf3ec74bd24569243798586f16a6ef7dffe330e98a54f22f094fe58735",
+      "size": 844
+    },
+    "source/results/20260916T183751Z-verify2b/manifest.json": {
+      "sha256": "e054c24da609c97b50fd3ecce33637bcc64f51c42ad723149bbbc2da786eeb70",
+      "size": 265
+    },
+    "source/results/20260916T183751Z-verify2b/reload_predictions.json": {
+      "sha256": "4e1ffbed45af3820bcf8c10d7638c925178e45284e9fc3cd86359ee230b80a67",
+      "size": 10453
+    },
+    "source/results/20260916T183751Z-verify2b/stress.json": {
+      "sha256": "efb4c0bc18cfca61511c7026a5b12e39fa423c00f0bfe63dd77343d03a52cbe5",
+      "size": 1163
+    },
+    "source/results/20260916T183751Z-verify2b/verification.json": {
+      "sha256": "49d95feaf1c72c076027b0aac71af26c80ecb6c7446755f88bd9171b79e6230c",
+      "size": 1920
+    },
+    "source/results/20260916T183823Z-train/correctness_final.json": {
+      "sha256": "5d6334a0c66ed567fa9709df4ea5b6040c88bb5258e864332e88be2bb9586634",
+      "size": 355
+    },
+    "source/results/20260916T183823Z-train/correctness_initial.json": {
+      "sha256": "6098996918d3fba844ad759459abb8063c2b7fdf9c65a891d7821a1d50dd9866",
+      "size": 423
+    },
+    "source/results/20260916T183823Z-train/data_filter.json": {
+      "sha256": "3a514e34a8a6f7e35776bafe5b19015d6c5cec0eaabea5558aa896471aba3b56",
+      "size": 2360
+    },
+    "source/results/20260916T183823Z-train/initial_validation_predictions.json": {
+      "sha256": "1189bd4f4d964b97e6fbeb8a10f8eaeb2c26e314e50b0a319b7ece322b325810",
+      "size": 324764
+    },
+    "source/results/20260916T183823Z-train/initial_validation_temperature_diagnostic.json": {
+      "sha256": "c466c9cb9b02206c5d1dab80db4fa579805f1ef890f549e6c27758240c35deda",
+      "size": 340
+    },
+    "source/results/20260916T183823Z-train/manifest.json": {
+      "sha256": "2cd93249a3c129203ddf3bb75827f810b30a61863e4dbf1bd5d2a1d7094419b8",
+      "size": 11460
+    },
+    "source/results/20260916T183823Z-train/summary.json": {
+      "sha256": "75044924b3f3b5a9b777627bccb1d80c7565b76a4520cc9c04228b0c7be93560",
+      "size": 11209
+    },
+    "source/results/20260916T183823Z-train/training.jsonl": {
+      "sha256": "2d425555511db013cae9ecafb28e2e1c7364c4dfcecd02b187f9c8c561a11e9e",
+      "size": 10422
+    },
+    "source/results/20260916T183823Z-train/validation.jsonl": {
+      "sha256": "964474c2ec7d6ee9c0f66055f51322944cc75f73a34adb5293849a3be6c1586c",
+      "size": 6173
+    },
+    "source/results/20260916T183823Z-train/validation_step_000040_predictions.json": {
+      "sha256": "e671e1508185765552b0f933ba03f356be62143c531d8ef534457d34b1645c9b",
+      "size": 333136
+    },
+    "source/results/20260916T185718Z-verify4b/data_filter.json": {
+      "sha256": "3a514e34a8a6f7e35776bafe5b19015d6c5cec0eaabea5558aa896471aba3b56",
+      "size": 2360
+    },
+    "source/results/20260916T185718Z-verify4b/http_255_choices_response.json": {
+      "sha256": "965b21465ccebb55e8c1b468da6e57653fe55dab5f2b48dec4ad207fdc322bfb",
+      "size": 12061
+    },
+    "source/results/20260916T185718Z-verify4b/http_long_context_response.json": {
+      "sha256": "fde2f6a2b43c93f77b292a5b89899254a7eee882598056a2a318d722c2b66f61",
+      "size": 216
+    },
+    "source/results/20260916T185718Z-verify4b/http_response.json": {
+      "sha256": "21ea4ea6c3e2c1c9e4383f5c5e340aaf6ee01b0903a9ce4d24672df63cf674d9",
+      "size": 861
+    },
+    "source/results/20260916T185718Z-verify4b/manifest.json": {
+      "sha256": "1fea5e941a3e38e11dc4e071c1dacb5de7c25345032207a95bb351e6cc211351",
+      "size": 265
+    },
+    "source/results/20260916T185718Z-verify4b/reload_predictions.json": {
+      "sha256": "2fee3e52111c3cd92babe0e36e5f2add009a4f69a3272a5fbacf7577026bc7f5",
+      "size": 10470
+    },
+    "source/results/20260916T185718Z-verify4b/stress.json": {
+      "sha256": "d93f33a15430685cc6357c9d2bdf08a10fa1e24d69d47fd58f66a4a9be7e46be",
+      "size": 1172
+    },
+    "source/results/20260916T185718Z-verify4b/verification.json": {
+      "sha256": "aede2c98c9773fd022e1ab807dfc66066a198dd01a9f35a6ca387e8e8341e48d",
+      "size": 2304
+    },
+    "source/results/20260916T185910Z-24h-launch/plan.json": {
+      "sha256": "979c1f0e0d4701c66115c209a25cf115d3210603a00d00ccd8b5ad027524b4fb",
+      "size": 649
+    },
+    "source/results/20260916T185910Z-24h-launch/resume_verification.json": {
+      "sha256": "772ab09829d7fb4403dcd7d3bb3685dee9e8559504b464191ff92f59a1a547d4",
+      "size": 933
+    },
+    "source/results/20260916T185910Z-24h-launch/state_snapshot.json": {
+      "sha256": "24cc3d6e9e17be546e68aea7654b4244d5b6bd883f5cf99ddb9e0c67cf87d7fa",
+      "size": 1328
+    },
+    "source/results/20260916T185910Z-24h-launch/training_correctness_initial.json": {
+      "sha256": "5d6334a0c66ed567fa9709df4ea5b6040c88bb5258e864332e88be2bb9586634",
+      "size": 355
+    },
+    "source/results/20260916T185910Z-24h-launch/training_manifest.json": {
+      "sha256": "b7c2e6e05d86f9c01a2aebd573e07ccd39da74bf25fd11e09d1150dc3326a619",
+      "size": 11618
+    },
+    "source/results/20260917-fleet-progress/ensemble-reference.json": {
+      "sha256": "3db8094bb4e01c2dfe74e880754b82c3ab08800529e357bdb4f1678beb21d40a",
+      "size": 6229
+    },
+    "source/results/20260917-fleet-progress/fixed-ensemble-validation.json": {
+      "sha256": "6c5dfa9d3528d4357cc89e90d10c7711b72eae50ddd67d67e15c4b902612c46d",
+      "size": 11460
+    },
+    "source/results/20260917-fleet-progress/fleet-four-candidates-status.json": {
+      "sha256": "2e2f2e6b8710562944b32c6d9b2619b415a6b79fd636dd85f1c1d1b48be186f4",
+      "size": 3234
+    },
+    "source/results/20260917-fleet-progress/four-candidate-registration.json": {
+      "sha256": "5d907228c8bc6ac48123359310bd5f56d567f2e72c30df3fd57d132c326f602d",
+      "size": 3568
+    },
+    "source/results/20260917-fleet-progress/gx10-status.json": {
+      "sha256": "3fb4d4c8d97d49d3f43460ff287bf8a619f863d4b93a9249b23b26f448b880a0",
+      "size": 4242
+    },
+    "source/results/20260917-fleet-progress/hf-final-watcher-launch.json": {
+      "sha256": "9654f1c752208de6c831fb7d6d9f1c0ec51c21f91351a148f9febd51aa579c41",
+      "size": 2519
+    },
+    "source/results/20260917-fleet-progress/hf-final-watcher-relaunch.json": {
+      "sha256": "a4e903e9eafeb4a911b0a532f8670b3b2b4f0cd9d324756009c078552a9c49d0",
+      "size": 1732
+    },
+    "source/results/20260917-fleet-progress/hf-initial-artifacts-publication.json": {
+      "sha256": "c8e938d34a13f17d3073c6ac4cb5de7b46f4cadf1e9f93f0e724766bf4f33e45",
+      "size": 2261
+    },
+    "source/results/20260917-fleet-progress/hf-snapshot-publication.json": {
+      "sha256": "686312f36485ada49c37bd2d1c129c39cdf86ff70316a87d65c0b0ef0cc363fc",
+      "size": 1031
+    },
+    "source/results/20260917-fleet-progress/hf-write-auth-verified.json": {
+      "sha256": "9e32210dd16078cf29f1339915f0712ba78925aadbe119d5a4c55b4f5512173f",
+      "size": 316
+    },
+    "source/results/20260917-fleet-progress/refinement-parent.json": {
+      "sha256": "f7c2765a5b9cf6fc794a30ec3b50bec986a046ed4deef080e3728b185c95a6ab",
+      "size": 1417
+    },
+    "source/results/20260917-fleet-progress/spark-a-status-20260917T021317Z.json": {
+      "sha256": "3c268f0e52bd5eb030f5d2799d51f6b33f52f750e7f277441c03ec83c24f93a4",
+      "size": 11252
+    },
+    "source/results/20260917-fleet-progress/spark-b-2b-completed-audit.json": {
+      "sha256": "53dea13fc44a987a071caa489ad2e832b0e434c516ada6de82423c5d2555aaef",
+      "size": 482109
+    },
+    "source/results/20260917-fleet-progress/spark-b-2b-completed-summary.json": {
+      "sha256": "16c665afdd0e3069ad2e18156da09c6dec8f58856b1ace9828feb559fd0e4620",
+      "size": 8395
+    },
+    "source/results/20260917-fleet-progress/spark-b-evidence-index.json": {
+      "sha256": "db281c7bfe8f023ed019038ab7c3e754fbdfda348a3eb5b2e9e4d90c73831b7c",
+      "size": 6475
+    },
+    "source/results/20260917-fleet-progress/spark-b-readonly-summary.json": {
+      "sha256": "16c665afdd0e3069ad2e18156da09c6dec8f58856b1ace9828feb559fd0e4620",
+      "size": 8395
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-best-validation-selection.json": {
+      "sha256": "a458754f2b605558fecc8c6506349d9cdb1b7d6bd25fb4f701d481b1e0a778bb",
+      "size": 19698
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-campaign-launch.json": {
+      "sha256": "203b65133d572f7b37bccfb6a4564ba5462efca651a797c3232d8c4420fede46",
+      "size": 4381
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-correctness-initial.json": {
+      "sha256": "f9ad8b0308a09577abc390412f214cc6289b1108ced3907570c6d4cf58539158",
+      "size": 354
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-current-status.json": {
+      "sha256": "a38095e1f9512022aad934a8db31d78985bfe0e336e4245c72203b6a5f68b7b2",
+      "size": 4735
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-data-filter.json": {
+      "sha256": "3a514e34a8a6f7e35776bafe5b19015d6c5cec0eaabea5558aa896471aba3b56",
+      "size": 2360
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-http-255-choices-response.json": {
+      "sha256": "b20caf2fb538c935bc5936c92c472082af58c54e4acd16dd2dc1447fe0b7d91f",
+      "size": 11995
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-http-long-context-response.json": {
+      "sha256": "8d13e7a618fa1a77a1cafead1afc6eee038411b3e09ff49d3e22afcbf6c0aa23",
+      "size": 215
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-http-response.json": {
+      "sha256": "cc339ccbd358dd410b52eaa9cb205971e8159eb3d2794ec2ee1a98f749943408",
+      "size": 864
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-inherited-validation-selection.json": {
+      "sha256": "a458754f2b605558fecc8c6506349d9cdb1b7d6bd25fb4f701d481b1e0a778bb",
+      "size": 19698
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-initial-predictions-verified.json": {
+      "sha256": "7817371003ac0883e10d6847fcd094cbc0f0ba9831ceedbb149e79aec9a23ccd",
+      "size": 707
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-initial-validation-selection.json": {
+      "sha256": "9675dfe39457a171238d81083d2c1e61dcb5aefa2429e2f2c9567ab79047e6c3",
+      "size": 19695
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-inputs-verified.json": {
+      "sha256": "8c5be28c56f988321f89dce076a1f6a834547795b2672b25d2d11ec60ad08bb4",
+      "size": 3472
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-pilot-correctness-final.json": {
+      "sha256": "f9ad8b0308a09577abc390412f214cc6289b1108ced3907570c6d4cf58539158",
+      "size": 354
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-pilot-correctness-initial.json": {
+      "sha256": "fd1b0d769f7a3f1cddacf00c19b0e3cec76c97c1d1e83a434ec2950bf178c677",
+      "size": 355
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-pilot-launch.json": {
+      "sha256": "d85c147ec85752b94eddc95b82ebdc6b93024d0915bdc0723ee65e941fde207d",
+      "size": 1419
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-pilot-manifest.json": {
+      "sha256": "e1ac6aaf86db2fff6c007a1018503877976ba04de7415125594142158b6f11a5",
+      "size": 12394
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-pilot-summary.json": {
+      "sha256": "929d4851d5b829a4cd69d337ab858f4f78480d6a654968d83141583c846fcec4",
+      "size": 11392
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-pilot-training.jsonl": {
+      "sha256": "6865669d51de01930927a807d94d34b78499004fbf6dfd78cba55ab26554ecb7",
+      "size": 2114
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-pilot-validation.jsonl": {
+      "sha256": "704b478c575080d30e1e555303b96b9307212d91c071673137f602a5b1db9fb6",
+      "size": 21145
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-pilot-verified.json": {
+      "sha256": "f39eebaabb44ff03e02353c93acdf7dafe512747b3778d4b4c553a7ceb7b78d4",
+      "size": 15161
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-plan.json": {
+      "sha256": "3e146d2daa7fa8e12689bebcfc8f3c8e27942328e15c2cf978a1950db21bb086",
+      "size": 777
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-reload-predictions.json": {
+      "sha256": "3c61f093082b06bf5274036d6e5f428465664ec9ae8125afc9ff99c781fc498a",
+      "size": 10396
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-resume-verified.json": {
+      "sha256": "0a2f072c250d91db2f419bd95c0c9eaee113229d3598f655d61ff05524773fcf",
+      "size": 3785
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-setup-status.json": {
+      "sha256": "c60e76f3a6a71b8893d1223db7091be667b3a95e02d9c919011ce4873b4b66e9",
+      "size": 199
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-startup-verified.json": {
+      "sha256": "97d75c66c1c1d76843cb2cb3a326780c452305501211da3b2ab3abef5a537b1b",
+      "size": 52576
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-stress.json": {
+      "sha256": "edc899aa2db8fae51a01818f0655a78b359b83dc521b3f7f399e458ffa5eeec7",
+      "size": 1166
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-training-manifest.json": {
+      "sha256": "0fe2c1bde4a47879d3ba040b97aa26c3025101dec10192e86188fe2cd98c33cf",
+      "size": 13002
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-verification-launch.json": {
+      "sha256": "ff87b649b1d2a9bb13d27e0bc8456d3fd5cd4076719e6833121ff69ca0775706",
+      "size": 1977
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-verification-manifest.json": {
+      "sha256": "997bee30fd9e1ea8420fb6671687b2e37c5aed76e6b1594521749bf73f35a01b",
+      "size": 264
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-verification.json": {
+      "sha256": "5d1a8d155bb53e27d2ffbe087d8b2e31686503fa277588c889d04865be3869c0",
+      "size": 2298
+    },
+    "source/results/20260917-fleet-progress/spark-b-refinement-warmstart-verified.json": {
+      "sha256": "72d8df546f492770738205561ecdc6c8c063b02c60f551ab752f3aafdebd78a7",
+      "size": 998
+    },
+    "source/results/20260917-fleet-progress/validation-trends.json": {
+      "sha256": "ddecc44c77276d0a6ca4d2418ad6f08a2e8caf00b327422e93f3407343610ac9",
+      "size": 4517
+    },
+    "source/results/20260917-playground/browser-verification.json": {
+      "sha256": "2806c5e75e560df7c0266e89aa92ac216bbca419be2094dee09e205d797ee460",
+      "size": 3596
+    },
+    "source/results/20260917-playground/concurrent-training.json": {
+      "sha256": "e581deac6f588ddf0a6effc595857a11bcbbf51979a5fdf812565e7611d608bc",
+      "size": 11763
+    },
+    "source/results/20260917-playground/desktop.png": {
+      "sha256": "0cbb5a6c9a9576ddceb95148024060350c5294779ed55063fb844f5978a3bf53",
+      "size": 138954
+    },
+    "source/results/20260917-playground/frontend-fixture-check.json": {
+      "sha256": "d368e15625343d2e94e3268a19e21fbd5f4a1cbe4994a4ef15b0542a83cfa1a6",
+      "size": 1179
+    },
+    "source/results/20260917-playground/launch.json": {
+      "sha256": "6af85b4328f8a253b3cc467290cad15eb0883268276eddeede0aa64d99395a54",
+      "size": 1991
+    },
+    "source/results/20260917-playground/mobile.png": {
+      "sha256": "e4711a2fa70b2d952e5f1b9e825dee5526c7eb6f837f99bf7f676aecd24c9af5",
+      "size": 127118
+    },
+    "source/results/20260917-playground/models.json": {
+      "sha256": "7a4e33a07801ab5fe918bb5a94c032900bd798513f6ab99608c9f06bb8c0572c",
+      "size": 1251
+    },
+    "source/results/20260917-playground/snapshots.json": {
+      "sha256": "a8187b1e1258b3c1ad50767c3e6b24d8e33b9533284f50347c8719c93ab1aea4",
+      "size": 1386
+    },
+    "source/results/checkpoint-sha256.txt": {
+      "sha256": "18fc16e95c38a28c1ec832f956c7ffc03fe9d0fdeb90348a84b73f2d9199937a",
+      "size": 254
+    },
+    "source/results/public-decisions-v1-manifest.json": {
+      "sha256": "adf5a8ca2bab60cf7429a82b1b7a2d1ae7d3cc812de04dd195a478dfd7c6f628",
+      "size": 5266
+    },
+    "source/scripts/campaign_status.py": {
+      "sha256": "1500b5e24f06231c7aafd7277aefd840582e35997e265f79db93614828d34411",
+      "size": 2301
+    },
+    "source/scripts/diagnose_parity.py": {
+      "sha256": "08b5d66d316ebda98a2226251a4f952701f86a1d5726ce7a4d7e8fb22755da5a",
+      "size": 3511
+    },
+    "source/scripts/download_candidate.py": {
+      "sha256": "d06a2c01be0cf6577f927fb37e3bc1eab014949fd934e4f4d9825674adba608e",
+      "size": 1141
+    },
+    "source/scripts/download_model.py": {
+      "sha256": "72ad9a5de44d09e2ee4ed8afb7c3c0ff6fb48987410a3f7f0368353572bf1f4d",
+      "size": 1013
+    },
+    "source/scripts/fleet_campaign.py": {
+      "sha256": "6f78b3df9ee3854330792d8218f91a4e907657e47bcca7f5d5e908044cbc5e78",
+      "size": 31181
+    },
+    "source/scripts/fleet_status.py": {
+      "sha256": "2519ced157ef4ac4fa449eebabae5740d7527778d578b4ac6720583010fa5217",
+      "size": 10289
+    },
+    "source/scripts/investigate_precision.py": {
+      "sha256": "609b744a926d8a45b87ba8d225e5312ee0c71a7096b21bd3589c5846e1dc847c",
+      "size": 10519
+    },
+    "source/scripts/launch_24h.py": {
+      "sha256": "39c26dc10535d3adac09209b0743edf2ed384563e732512ad62f8c6f37161a83",
+      "size": 19127
+    },
+    "source/scripts/prepare_public_data.py": {
+      "sha256": "32ea84aa719818e1141b958b6ef27a85f7ddb86bfcd7c1585fc487b25253d977",
+      "size": 11886
+    },
+    "source/scripts/publish_hf_final.py": {
+      "sha256": "278efc5878d7ebc5d1171f3a735d6c7d78275d93650e5a8a16c3f55b8353a577",
+      "size": 20356
+    },
+    "source/scripts/publish_hf_snapshot.py": {
+      "sha256": "b5fe16a00fcbc5ab97121428c6ce750275ee193438e3ffe24aac5e3bb325c018",
+      "size": 15165
+    },
+    "source/scripts/run_experiment.sh": {
+      "sha256": "661a6309fc54a2a8aff918f14a553c72dcb21730bd6a3cfc55d6ccd4700d11d3",
+      "size": 1213
+    },
+    "source/scripts/run_precision.sh": {
+      "sha256": "766d82b30cf3e83951f662685b4472ee053c7fbc171c4ff140823bf4d1c2782f",
+      "size": 998
+    },
+    "source/scripts/run_smoke.sh": {
+      "sha256": "39d59f2120f362729d1c2e384391b82be1e580dcc1aca0eaa7ab231115225574",
+      "size": 1054
+    },
+    "source/scripts/start_spark_candidate.sh": {
+      "sha256": "c2ca18b008a144de7cb264c9fcca634d68c4a8a317db3638e8a70b3dbcff064b",
+      "size": 6545
+    },
+    "source/scripts/verify_artifact.py": {
+      "sha256": "9833350e9d72c0065b15206bb71c5a8b5a6b3185219db90074e369ede563a985",
+      "size": 9971
+    },
+    "source/scripts/verify_playground.cjs": {
+      "sha256": "0be5180bf1d2f812a7e416099b3b6263ef3fa766b2b1057418495eb9592fa48e",
+      "size": 5866
+    },
+    "source/selection.py": {
+      "sha256": "be0a7a8496b5b830aa572ceba93606320f442063fd38180503fd6980dc1c578f",
+      "size": 5335
+    },
+    "source/smoke_data.py": {
+      "sha256": "06b3cbac1c8c4a86b8aecbee4459073cc3e46d4ddcd576392f3cb4805924f815",
+      "size": 1807
+    },
+    "source/smoke_train.py": {
+      "sha256": "8cdeb2b397177fc9c26638aaa871501ddab8e3871aa1573ecd98f66590f5c228",
+      "size": 19630
+    },
+    "source/tests/test_campaign.py": {
+      "sha256": "90b132f655db9e9fd7b71c4916d47064c51e65cc1cc0de1a69f0b0cdc8d1cad7",
+      "size": 8567
+    },
+    "source/tests/test_fleet_campaign.py": {
+      "sha256": "ed5fc03bdc5d5c3de28b1069aadb93e9ca1d30965e412bcafc43dd97d86724e3",
+      "size": 15318
+    },
+    "source/tests/test_fleet_status.py": {
+      "sha256": "681eabe15b1de8380302a15f87cba411037da44c73a5e8b817455a7bb13dd061",
+      "size": 3469
+    },
+    "source/tests/test_playground.py": {
+      "sha256": "f63b58eef3b0f9c42aa3d445ceb4fa5462e936deb7b11e2e0fc962af6ed32c96",
+      "size": 10663
+    },
+    "source/tests/test_publish_hf_final.py": {
+      "sha256": "dd18fb43526d23e36166787a1d3eae5a12aedd6ffc54951459c97c37b1f2c3ee",
+      "size": 15397
+    },
+    "source/tests/test_publish_hf_snapshot.py": {
+      "sha256": "f48905828b51060f9a265505245eb10f814b47b8610ce5b01d67b2005c828edd",
+      "size": 9332
+    },
+    "source/tests/test_scorer.py": {
+      "sha256": "2c6f5d5e9ff634049cbe9c88b5f126710298b65ce603d7ced2072bc06e3974d8",
+      "size": 5091
+    },
+    "source/tests/test_selection.py": {
+      "sha256": "09fc10fd394ea870307175698a81b516a0bafc536757be918f5cacb86d124770",
+      "size": 4090
+    },
+    "source/tests/test_training_harness.py": {
+      "sha256": "df8175b7d051625142844cece4b1bf709f395a39f64319c8eeb4f6edd8140b87",
+      "size": 22141
+    },
+    "source/training_model.py": {
+      "sha256": "d5b0aefeeb5290816bc0b669aa0a8cbbe27f6a12b9cb23c141ac9b9ae9ee4e65",
+      "size": 7862
+    },
+    "source/web/playground.css": {
+      "sha256": "09969ed5332298048595879c930d1065850db3c915688ea2da343845e879f27b",
+      "size": 14088
+    },
+    "source/web/playground.html": {
+      "sha256": "69495feec84f29a6efd06c7e45c1ddac3d37717a16b7b9036093609dfd24b4df",
+      "size": 8847
+    },
+    "source/web/playground.js": {
+      "sha256": "b2ccc2dab0fb670fb3cd11e43f641cdab7c3752724351e7641f4dd3a4ca86646",
+      "size": 19280
+    },
+    "sources/4a60423c39d70f8d50472ce4f4f7fa4a4bd9fce1/manifest.json": {
+      "sha256": "44f77f5d586ff77ce6cb7ddffe901d1f93b1afb4aeef322024d7de209faf0771",
+      "size": 23766
+    },
+    "sources/4a60423c39d70f8d50472ce4f4f7fa4a4bd9fce1/source.tar.gz": {
+      "sha256": "6f76ba9b5ee699249f5258d52ef05f372aea3f425dc951c2681b1b4a1f2be918",
+      "size": 700511
+    },
+    "sources/b05f7e3975a974c26ab667c6fdf91ec489a8cec0/manifest.json": {
+      "sha256": "68ad005033485833adbd1a16f5ca8c0f3ebdce65a2638a145e36ea05eb6cb33c",
+      "size": 48983
+    },
+    "sources/b05f7e3975a974c26ab667c6fdf91ec489a8cec0/source.tar.gz": {
+      "sha256": "64a88e5e218f68e4b1dda00e7012eee0aaba4e4fcb9e0443b081c188528946c9",
+      "size": 1654578
+    }
+  },
+  "format": "opensysone-snapshot-publication-v1",
+  "repo_id": "andyshu/opensysone",
+  "snapshot_id": "20260917T022835Z-hf-backup",
+  "snapshot_path": "snapshots/20260917T022835Z-hf-backup",
+  "source_commit": "b05f7e3975a974c26ab667c6fdf91ec489a8cec0",
+  "training_source_commits": [
+    "4a60423c39d70f8d50472ce4f4f7fa4a4bd9fce1"
+  ]
+}

source/HANDOVER.md

@@ -268,3 +268,43 @@ with the process above to remove inherited `HF_TOKEN`/`HUGGING_FACE_HUB_TOKEN`
 overrides. It will read the updated write-capable stored login when final publication
 begins. Changing credentials does not require
 changing the training jobs, fleet plan, API or repository visibility.
+
+
+## Interactive model playground — 2026-09-17
+
+The user requested a GUI for text plus candidate answers and probabilities. It is
+running on GX10 at **http://127.0.0.1:18082**, PID **1467162**, source **`e94e58e`**,
+OOM adjustment 0. From the Mac, run `ssh -N -L 18082:127.0.0.1:18082 gx10`, then
+open **http://localhost:18082**. See [PLAYGROUND.md](PLAYGROUND.md).
+
+Runtime: `/home/andy/ai/opensysone/runs/20260917T032753Z-playground`, also recorded
+in `LAST_PLAYGROUND`. `launch.json` records the exact process command/source
+hashes and `server.log` receives sanitized diagnostics. Exit status is pending
+while serving. Inspect `/api/status` and match `/proc/1467162/cmdline` against
+`launch.json.command` before sending SIGTERM to this process only. The documented
+CLI restarts it from the fixed catalog after the old listener has stopped.
+
+Three immutable snapshots are available: GX10 4B step 2,500, Spark A 4B step 2,500,
+and Spark B 2B step 2,000. Their files/hashes and matching provenance are under
+this runtime's `models/`; these probabilities are explicitly uncalibrated. No
+reserved evaluation examples were used for GUI testing. One backend resides at
+a time; loads, scoring and unloading are serialized on a dedicated worker thread.
+The 16 GiB allocation cap and memory/OOM checks remain active. This extra GUI
+process shares GPU compute with training, so requests can slow optimizer steps.
+Port 18081 remains reserved for final deployment; no firewall/services changed.
+
+All seven backend tests passed. Chromium passed real inference for all three
+models, return switching, clipboard JSON, input-edit staleness, duplicate options,
+actual tokenizer overflow and mobile layout. Browser script/CSP errors: none.
+Cold/switch example requests were 5.08–9.10 seconds; a warm main-model request
+was 1.53 seconds. These are individual wiring timings, not latency percentiles
+or quality estimates. Real results/screenshots and concurrency observations are
+in `results/20260917-playground/`; the separate frontend fixture report is labeled
+as stubbed UI testing. Training and the fleet/final-publication controllers remain
+independent of this GUI.
+
+A 45-second observation after GUI verification recorded the GX10 trainer advancing
+from step 3,063 to 3,068 with finite losses/gradients, a 15.624 GiB allocation peak
+and 81.3 GiB host memory available. The GUI stayed ready. This confirms continued
+training during GUI operation; it does not establish zero slowdown or capture all
+model-switch transients.

source/HF_MODEL_CARD.md

@@ -50,6 +50,12 @@ why independent trials were retained over unmeasured distributed training.
 
 ## Contents and reconstruction
 
+The [browser playground](source/PLAYGROUND.md) provides a context box, question,
+options list and probability bars, with a model selector. Its lightweight local
+server uses these custom checkpoints and the pinned base weights. It runs on
+GX10 over a loopback connection/SSH tunnel; this repository is a backup, not a
+hosted inference application.
+
 - `source/`: a committed source/documentation snapshot, including the training
   scripts, tests, Jev-compatible harness, findings and operational handover.
 - `snapshots/<id>/backup-manifest.json`: artifact roles, originating hosts/paths,

source/JEV_HARNESS.md

@@ -1,5 +1,9 @@
 # Use OpenSysOne and hosted Jev with the same request
 
+For an interactive text-and-options interface, use the
+[browser playground](PLAYGROUND.md). It serves frozen training snapshots through
+this scoring code while the longer training campaign continues.
+
 The harness implements TypeSafe's documented `POST /v1/systemone` request and
 answer shapes for `choice`, `score` and `noul`. See the
 [official API reference](https://docs.typesafe.ai/api) and

source/PLAYGROUND.md

@@ -0,0 +1,102 @@
+# Try the trained models
+
+The playground accepts context, a question and one option per line, then shows a
+probability for each option. It uses real frozen OpenSysOne checkpoints through
+the existing scoring harness. Pick another model to try the same decision again.
+
+On your Mac, leave this tunnel running:
+
+```bash
+ssh -N -L 18082:127.0.0.1:18082 gx10
+```
+
+Open **http://localhost:18082**. If you are using a browser directly on GX10, no
+tunnel is needed. Port 18081 remains reserved for the final evaluated model API.
+
+Paste your text, adjust the question and enter at least two distinct options.
+Click **Get probabilities**, or press **Command/Ctrl + Enter**. The examples are
+invented inputs you can edit. Copy JSON exports the actual result to your clipboard.
+Editing an input invalidates the previous result so it cannot be confused with a
+new question. Inputs are not saved to browser storage or request logs.
+
+The available snapshots are:
+
+| Model | Selected step | Calibration |
+| --- | ---: | --- |
+| Qwen3 4B · Main | 2,500 | Uncalibrated |
+| Qwen3 4B · Lower rate | 2,500 | Uncalibrated |
+| Qwen3.5 2B | 2,000 | Uncalibrated |
+
+Probabilities are normalized over the supplied options. Adding/removing an option
+changes the question being scored. An uncalibrated 90% output is not a demonstrated
+90% success rate. These fixed snapshots remain stable while training continues.
+
+Each complete context/question/option prompt must fit the verified **1,024-token**
+inference limit, including chat formatting. The server reports an error for longer
+inputs instead of silently truncating them. Up to 255 distinct options are allowed;
+large lists take longer. The first request after selecting another model also
+loads its weights. One request runs at a time; other requests receive a busy reply.
+
+## Runtime and controls
+
+The current server is running as PID **1467162**, source **`e94e58e`**, with pending
+exit status. Its immutable runtime is
+`/home/andy/ai/opensysone/runs/20260917T032753Z-playground`.
+
+The runtime directory is recorded in `~/ai/opensysone/runs/LAST_PLAYGROUND`.
+Its `models.json` lists the fixed snapshots, and its launch/verification evidence
+records the process, source revision, log and numerical checks. `HANDOVER.md`
+records the active process. Inspect the listener and status with:
+
+```bash
+curl --fail http://127.0.0.1:18082/api/status
+curl --fail http://127.0.0.1:18082/api/models
+```
+
+Use the existing isolated environment from this project directory to start it:
+
+```bash
+~/ai/envs/opensysone/bin/python playground.py \
+  --models /home/andy/ai/opensysone/runs/20260917T032753Z-playground/models.json \
+  --port 18082 --device cuda --max-tokens 1024
+```
+
+Before restarting, verify and stop the exact recorded playground process with
+SIGTERM. This is separate from the training jobs, fleet coordinator and final-model
+uploader. The server binds only to loopback; it requires no firewall or service
+configuration changes. No frontend dependency installation is required.
+
+The server loads one model at a time, checks frozen checkpoint hashes and model
+metadata, and releases a previous model before loading a different one. The same
+24 GiB available-memory check, 16 GiB CUDA allocation cap, FP32 scoring and OOM
+adjustment 0 apply. GPU inference shares compute with GX10 training; frequent or
+large requests can slow that training. Idle browser tabs perform no model inference.
+The runtime supports `--device cpu` as an alternative, with latency depending on
+the model and input. Check current memory and GPU processes before any model load.
+
+The frontend is in `web/`, the server is `playground.py`, and the input/scoring
+contract is inherited from `jev_harness.py`. See [JEV_HARNESS.md](JEV_HARNESS.md)
+for the underlying API and hosted Jev integration.
+
+
+## Verification
+
+Seven backend tests pass with
+`python3 -m unittest discover -s tests -p 'test_playground.py'`.
+Real Chromium checks and screenshots are in
+[`results/20260917-playground`](results/20260917-playground/). They score invented
+examples with all three actual models; reserved test/holdout data remains untouched.
+The main-model warm example took 1.53 seconds; first-load/model-switch requests
+were 5.08–9.10 seconds. These single observations are not latency percentiles.
+
+To repeat the real browser check using existing GX10 browser tooling:
+
+```bash
+PLAYWRIGHT_MODULE=/home/andy/projects/ultraviolet/node_modules/playwright-core \
+CHROMIUM_PATH=/home/andy/.cache/ms-playwright/chromium-1234/chrome-linux/chrome \
+  node scripts/verify_playground.cjs http://127.0.0.1:18082 \
+  /home/andy/ai/opensysone/runs/NEW-UNUSED-playground-browser-check
+```
+
+That command performs real inference and changes the resident GUI model during
+its checks. It does not train or alter checkpoints.

source/README.md

@@ -1,5 +1,8 @@
 # OpenSysOne
 
+Try the trained models in the [browser playground](PLAYGROUND.md): enter context,
+a question and possible answers, then compare their probabilities across models.
+
 Experimental decision scorer: pretrained Transformer, arbitrary natural-language
 candidates and scalar logits, with a Jev-compatible API harness.
 

source/playground.py

@@ -0,0 +1,365 @@
+"""Loopback-only browser playground for a fixed catalog of trusted checkpoints."""
+import argparse
+from concurrent.futures import ThreadPoolExecutor
+import gc
+import hashlib
+from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
+import json
+from pathlib import Path
+import re
+import signal
+import subprocess
+import sys
+import threading
+import time
+from urllib.parse import urlsplit
+
+ROOT = Path(__file__).resolve().parent
+MAX_BODY_BYTES = 128 * 1024
+ASSETS = {'/': ('playground.html', 'text/html; charset=utf-8'),
+          '/playground.html': ('playground.html', 'text/html; charset=utf-8'),
+          '/playground.css': ('playground.css', 'text/css; charset=utf-8'),
+          '/playground.js': ('playground.js', 'text/javascript; charset=utf-8')}
+CSP = ("default-src 'none'; script-src 'self'; style-src 'self'; connect-src 'self'; "
+       "img-src 'self' data:; base-uri 'none'; frame-ancestors 'none'; form-action 'none'")
+
+
+class RequestError(Exception):
+    def __init__(self, status, code, message):
+        super().__init__(message)
+        self.status, self.code, self.message = status, code, message
+
+
+def checksum(path):
+    value = hashlib.sha256()
+    with Path(path).open('rb') as handle:
+        for chunk in iter(lambda: handle.read(1024 * 1024), b''):
+            value.update(chunk)
+    return value.hexdigest()
+
+
+def load_catalog(path, max_tokens=1024):
+    if not isinstance(max_tokens, int) or isinstance(max_tokens, bool) or not 1 <= max_tokens <= 32768:
+        raise ValueError('Invalid inference token limit')
+    data = json.loads(Path(path).read_text())
+    models = data.get('models')
+    if not isinstance(models, list) or not models:
+        raise ValueError('Model catalog must contain a nonempty models list')
+    catalog = {}
+    for model in models:
+        if not isinstance(model, dict) or not re.fullmatch(r'[A-Za-z0-9][A-Za-z0-9_.-]{0,63}', model.get('id', '')):
+            raise ValueError('Invalid catalog model ID')
+        if model['id'] in catalog:
+            raise ValueError('Duplicate catalog model ID')
+        if any(not isinstance(model.get(key), str) or not model[key].strip() for key in ('label', 'description', 'checkpoint')):
+            raise ValueError('Catalog needs a label, description and checkpoint')
+        if type(model.get('calibrated')) is not bool or type(model.get('checkpoint_step')) is not int or model['checkpoint_step'] < 0:
+            raise ValueError('Catalog needs explicit calibration and checkpoint-step metadata')
+        checkpoint = Path(model['checkpoint'])
+        if not checkpoint.is_absolute() or checkpoint.is_symlink() or not checkpoint.is_file():
+            raise ValueError('Catalog checkpoint must be an absolute regular file')
+        actual_hash = checksum(checkpoint)
+        if model.get('sha256', actual_hash) != actual_hash:
+            raise ValueError('Catalog checkpoint checksum mismatch')
+        catalog[model['id']] = {key: model[key] for key in ('id', 'label', 'description', 'calibrated', 'checkpoint_step')}
+        catalog[model['id']].update(checkpoint=str(checkpoint.resolve()), sha256=actual_hash, max_tokens=max_tokens)
+    default = data.get('default_model', next(iter(catalog)))
+    if default not in catalog:
+        raise ValueError('Default model is absent from catalog')
+    return catalog, default
+
+
+def local_backend(checkpoint, device, max_tokens):
+    from jev_harness import LocalBackend
+    return LocalBackend(checkpoint, device=device, max_tokens=max_tokens)
+
+
+def memory_preflight(device):
+    Path('/proc/self/oom_score_adj').write_text('0')
+    subprocess.run(['free', '-b'], check=True, capture_output=True, text=True, timeout=10)
+    available = int(next(line.split()[1] for line in Path('/proc/meminfo').read_text().splitlines()
+                         if line.startswith('MemAvailable:'))) * 1024
+    if device == 'cuda':
+        # Inspect all GPU allocations; coexistence is allowed when RAM is sufficient.
+        subprocess.run(['nvidia-smi', '--query-compute-apps=pid,process_name,used_memory', '--format=csv'],
+                       check=True, capture_output=True, text=True, timeout=10)
+    if available < 24 * 2**30:
+        raise RequestError(503, 'low_memory', 'At least 24 GiB of available system memory is required to load a model.')
+
+
+def release_memory(device):
+    gc.collect()
+    torch = sys.modules.get('torch')
+    if device == 'cuda' and torch is not None and torch.cuda.is_initialized():
+        torch.cuda.empty_cache()
+
+
+def release_exception_frames(error):
+    # Future/HTTP error objects must not retain failed scorer frames or tensors.
+    error.__traceback__ = None
+    error.__context__ = None
+    error.__cause__ = None
+
+
+class ModelService:
+    def __init__(self, catalog, default_model, device='cuda', backend_factory=local_backend,
+                 preflight=memory_preflight, cleanup=release_memory):
+        self.catalog, self.default_model, self.device = catalog, default_model, device
+        self.backend_factory, self.preflight, self.cleanup = backend_factory, preflight, cleanup
+        self.lock, self.state_lock = threading.Lock(), threading.Lock()
+        self.executor = ThreadPoolExecutor(max_workers=1, thread_name_prefix='playground-model')
+        self.backend = None
+        self.state = {'status': 'idle', 'loaded_model_id': None}
+
+    def status(self):
+        with self.state_lock:
+            return dict(self.state)
+
+    def set_state(self, status, **values):
+        with self.state_lock:
+            self.state.update(status=status, **values)
+
+    def models(self):
+        keys = ('id', 'label', 'description', 'calibrated', 'max_tokens', 'checkpoint_step')
+        return {'models': [{key: model[key] for key in keys} for model in self.catalog.values()],
+                'default_model': self.default_model}
+
+    def unload(self):
+        self.backend = None
+        self.set_state('loading', loaded_model_id=None)
+        self.cleanup(self.device)
+
+    def validate(self, request):
+        if not isinstance(request, dict) or set(request) != {'model', 'text', 'question', 'options'}:
+            raise RequestError(422, 'invalid_request', 'Provide model, text, question and options.')
+        if not isinstance(request['model'], str) or request['model'] not in self.catalog:
+            raise RequestError(422, 'unknown_model', 'Choose a model from the configured catalog.')
+        for key in ('text', 'question'):
+            if not isinstance(request[key], str) or not request[key].strip():
+                raise RequestError(422, 'invalid_request', 'Context and question must be nonempty text.')
+        options = request['options']
+        if (not isinstance(options, list) or not 2 <= len(options) <= 255 or
+                any(not isinstance(option, str) or not option.strip() for option in options) or
+                len(set(options)) != len(options)):
+            raise RequestError(422, 'invalid_options', 'Provide 2–255 distinct, nonempty answer options.')
+
+    def score(self, request):
+        self.validate(request)
+        if not self.lock.acquire(blocking=False):
+            raise RequestError(503, 'busy', 'A model is loading or scoring. Try again when it finishes.')
+        try:
+            pending = self.executor.submit(self._score, request)
+        except BaseException:
+            self.lock.release()
+            raise
+        return pending.result()
+
+    def close(self):
+        self.executor.submit(self.unload).result()
+        self.executor.shutdown(wait=True)
+        self.set_state('idle')
+
+    def _score(self, request):
+        started = time.perf_counter()
+        model = self.catalog[request['model']]
+        try:
+            if self.status()['loaded_model_id'] != model['id']:
+                self.unload()
+                try:
+                    self.preflight(self.device)
+                    if checksum(model['checkpoint']) != model['sha256']:
+                        raise ValueError('Configured artifact changed')
+                    self.backend = self.backend_factory(model['checkpoint'], device=self.device, max_tokens=model['max_tokens'])
+                    if self.backend.calibrated is not model['calibrated'] or self.backend.max_tokens != model['max_tokens']:
+                        raise ValueError('Loaded artifact does not match catalog metadata')
+                except Exception as error:
+                    if not isinstance(error, RequestError):
+                        print(json.dumps({'event': 'model_load_failed', 'model_id': model['id'],
+                                          'error_type': type(error).__name__}), flush=True)
+                    # Drop construction tracebacks before releasing partial allocations.
+                    release_exception_frames(error)
+                    self.unload()
+                    self.set_state('error')
+                    if isinstance(error, RequestError):
+                        raise error from None
+                    raise RequestError(503, 'model_unavailable', 'The configured model could not be loaded.') from None
+                self.set_state('ready', loaded_model_id=model['id'])
+            payload = {'model': 'opensysone', 'state': request['text'], 'questions': {'answer': {
+                'type': 'choice', 'instructions': request['question'],
+                'criteria': {option: None for option in request['options']}}}}
+            self.set_state('scoring')
+            try:
+                result = self.backend(payload)
+                from jev_harness import validate_response
+                validate_response(payload, result)
+                probabilities = result['answers']['answer']['probabilities']
+                tokens = result['usage']['input_tokens']
+                if type(tokens) is not int or tokens < 0:
+                    raise ValueError('Invalid backend token count')
+            except ValueError as error:
+                self.set_state('error')
+                match = re.fullmatch(r'Input has (\d+) tokens; limit is (\d+); no truncation', str(error))
+                release_exception_frames(error)
+                if match:
+                    raise RequestError(422, 'token_limit', f'Input has {match[1]} tokens; the limit is {match[2]}. Shorten the text, question or options; nothing was truncated.') from None
+                print(json.dumps({'event': 'model_score_failed', 'model_id': model['id'],
+                                  'error_type': type(error).__name__}), flush=True)
+                raise RequestError(500, 'scoring_failed', 'The model could not produce a valid result.') from None
+            except Exception as error:
+                self.set_state('error')
+                release_exception_frames(error)
+                print(json.dumps({'event': 'model_score_failed', 'model_id': model['id'],
+                                  'error_type': type(error).__name__}), flush=True)
+                raise RequestError(500, 'scoring_failed', 'The model could not score this request.') from None
+            self.set_state('ready')
+            return {'model': {key: model[key] for key in ('id', 'label', 'calibrated', 'checkpoint_step')},
+                    'probabilities': [{'option': option, 'probability': probabilities[option]} for option in request['options']],
+                    'elapsed_seconds': time.perf_counter() - started, 'input_tokens': tokens}
+        finally:
+            self.lock.release()
+
+
+def loopback_authority(value):
+    try:
+        parsed = urlsplit('//' + value)
+        if (parsed.hostname not in ('127.0.0.1', 'localhost', '::1') or parsed.username is not None or
+                parsed.password is not None or parsed.path or parsed.query or parsed.fragment):
+            return None
+        return parsed.hostname, parsed.port or 80
+    except (ValueError, TypeError):
+        return None
+
+
+def create_server(service, port=18082, web_root=ROOT / 'web'):
+    web_root = Path(web_root)
+    class Handler(BaseHTTPRequestHandler):
+        def setup(self):
+            super().setup()
+            self.connection.settimeout(15)
+
+        def send_data(self, status, content, content_type):
+            self.send_response(status)
+            for key, value in (('Content-Type', content_type), ('Content-Length', str(len(content))),
+                               ('Cache-Control', 'no-store'), ('X-Content-Type-Options', 'nosniff'),
+                               ('Content-Security-Policy', CSP), ('Referrer-Policy', 'no-referrer')):
+                self.send_header(key, value)
+            self.end_headers()
+            try:
+                self.wfile.write(content)
+            except (BrokenPipeError, ConnectionResetError):
+                pass
+
+        def json(self, status, value):
+            self.send_data(status, json.dumps(value, allow_nan=False).encode(), 'application/json; charset=utf-8')
+
+        def error(self, error):
+            self.json(error.status, {'error': error.message, 'code': error.code})
+
+        def send_error(self, code, message=None, explain=None):
+            self.json(code, {'error': 'HTTP request rejected.', 'code': 'http_error'})
+
+        def check_origin(self, post=False):
+            hosts = self.headers.get_all('Host', [])
+            authority = loopback_authority(hosts[0]) if len(hosts) == 1 else None
+            if authority is None:
+                raise RequestError(403, 'forbidden_origin', 'Use a loopback address or local SSH tunnel.')
+            if post:
+                origins = self.headers.get_all('Origin', [])
+                if len(origins) > 1:
+                    raise RequestError(403, 'forbidden_origin', 'The request must come from this playground.')
+                if origins:
+                    try:
+                        parsed = urlsplit(origins[0])
+                    except ValueError:
+                        raise RequestError(403, 'forbidden_origin', 'The request must come from this playground.') from None
+                    if (parsed.scheme not in ('http', 'https') or parsed.path or parsed.query or parsed.fragment or
+                            loopback_authority(parsed.netloc) != authority):
+                        raise RequestError(403, 'forbidden_origin', 'The request must come from this playground.')
+                if self.headers.get('Sec-Fetch-Site') in ('cross-site', 'same-site'):
+                    raise RequestError(403, 'forbidden_origin', 'The request must come from this playground.')
+
+        def do_GET(self):
+            try:
+                self.check_origin()
+                path = urlsplit(self.path).path
+                if path == '/api/models':
+                    return self.json(200, service.models())
+                if path == '/api/status':
+                    return self.json(200, service.status())
+                if path not in ASSETS:
+                    raise RequestError(404, 'not_found', 'Not found.')
+                filename, content_type = ASSETS[path]
+                try:
+                    content = (web_root / filename).read_bytes()
+                except OSError:
+                    raise RequestError(404, 'not_found', 'Playground asset is unavailable.') from None
+                self.send_data(200, content, content_type)
+            except RequestError as error:
+                self.error(error)
+            except Exception:
+                self.error(RequestError(500, 'internal_error', 'The request could not be completed.'))
+
+        def do_POST(self):
+            try:
+                self.check_origin(post=True)
+                if urlsplit(self.path).path != '/api/score':
+                    raise RequestError(404, 'not_found', 'Not found.')
+                if self.headers.get_content_type() != 'application/json':
+                    raise RequestError(415, 'content_type', 'Use application/json.')
+                lengths = self.headers.get_all('Content-Length', [])
+                if self.headers.get('Transfer-Encoding') or len(lengths) != 1 or not lengths[0].isdigit():
+                    raise RequestError(400, 'invalid_length', 'A valid Content-Length is required.')
+                length = int(lengths[0])
+                if not 0 < length <= MAX_BODY_BYTES:
+                    raise RequestError(413, 'body_too_large', 'Request body must be 1 byte to 128 KiB.')
+                raw = self.rfile.read(length)
+                if len(raw) != length:
+                    raise RequestError(400, 'invalid_body', 'Request body was incomplete.')
+                try:
+                    request = json.loads(raw)
+                except (ValueError, UnicodeError):
+                    raise RequestError(400, 'invalid_json', 'Request body must be valid JSON.') from None
+                self.json(200, service.score(request))
+            except RequestError as error:
+                self.error(error)
+            except (TimeoutError, OSError):
+                self.error(RequestError(408, 'request_timeout', 'The request could not be read in time.'))
+            except Exception:
+                self.error(RequestError(500, 'internal_error', 'The request could not be completed.'))
+
+        def log_message(self, format, *args):
+            # Do not persist URLs, bodies, prompt text, or request headers.
+            pass
+
+    server = ThreadingHTTPServer(('127.0.0.1', port), Handler)
+    server.daemon_threads = True
+    return server
+
+
+def main():
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument('--models', required=True, help='Trusted server-side JSON model catalog')
+    parser.add_argument('--port', type=int, default=18082)
+    parser.add_argument('--device', choices=('cuda', 'cpu'), default='cuda')
+    parser.add_argument('--max-tokens', type=int, default=1024)
+    args = parser.parse_args()
+    if not 1 <= args.port <= 65535:
+        parser.error('Port must be between 1 and 65535')
+    Path('/proc/self/oom_score_adj').write_text('0')
+    catalog, default = load_catalog(args.models, args.max_tokens)
+    service = ModelService(catalog, default, args.device)
+    server = create_server(service, args.port)
+    signal.signal(signal.SIGTERM, lambda signum, frame: threading.Thread(target=server.shutdown, daemon=True).start())
+    print(json.dumps({'event': 'playground_listening', 'host': '127.0.0.1', 'port': args.port,
+                      'model_ids': list(catalog)}), flush=True)
+    try:
+        server.serve_forever()
+    except KeyboardInterrupt:
+        pass
+    finally:
+        server.server_close()
+        service.close()
+
+
+if __name__ == '__main__':
+    main()

source/results/20260917-playground/browser-verification.json

@@ -0,0 +1,137 @@
+{
+  "started_utc": "2026-09-17T03:34:55.355Z",
+  "url": "http://127.0.0.1:18082",
+  "real_model_responses": [
+    {
+      "model": {
+        "id": "gx10-4b",
+        "label": "Qwen3 4B · Main",
+        "calibrated": false,
+        "checkpoint_step": 2500
+      },
+      "probabilities": [
+        {
+          "option": "Visit the free art gallery",
+          "probability": 1
+        },
+        {
+          "option": "Take a long hike",
+          "probability": 3.0575177806246856e-11
+        },
+        {
+          "option": "Book a rooftop lunch",
+          "probability": 1.0092975066822163e-11
+        }
+      ],
+      "elapsed_seconds": 9.100128590973327,
+      "input_tokens": 391
+    },
+    {
+      "model": {
+        "id": "spark-a-4b",
+        "label": "Qwen3 4B · Lower rate",
+        "calibrated": false,
+        "checkpoint_step": 2500
+      },
+      "probabilities": [
+        {
+          "option": "Visit the free art gallery",
+          "probability": 1
+        },
+        {
+          "option": "Take a long hike",
+          "probability": 8.72549676955714e-9
+        },
+        {
+          "option": "Book a rooftop lunch",
+          "probability": 1.7202640423263915e-9
+        }
+      ],
+      "elapsed_seconds": 7.556980039022164,
+      "input_tokens": 391
+    },
+    {
+      "model": {
+        "id": "spark-b-2b",
+        "label": "Qwen3.5 2B",
+        "calibrated": false,
+        "checkpoint_step": 2000
+      },
+      "probabilities": [
+        {
+          "option": "Visit the free art gallery",
+          "probability": 0.9999608993530273
+        },
+        {
+          "option": "Take a long hike",
+          "probability": 0.00002338631202292163
+        },
+        {
+          "option": "Book a rooftop lunch",
+          "probability": 0.00001569445157656446
+        }
+      ],
+      "elapsed_seconds": 5.078740055003436,
+      "input_tokens": 418
+    },
+    {
+      "model": {
+        "id": "gx10-4b",
+        "label": "Qwen3 4B · Main",
+        "calibrated": false,
+        "checkpoint_step": 2500
+      },
+      "probabilities": [
+        {
+          "option": "Visit the free art gallery",
+          "probability": 1
+        },
+        {
+          "option": "Take a long hike",
+          "probability": 3.0575177806246856e-11
+        },
+        {
+          "option": "Book a rooftop lunch",
+          "probability": 1.0092975066822163e-11
+        }
+      ],
+      "elapsed_seconds": 6.513889841007767,
+      "input_tokens": 391
+    },
+    {
+      "model": {
+        "id": "gx10-4b",
+        "label": "Qwen3 4B · Main",
+        "calibrated": false,
+        "checkpoint_step": 2500
+      },
+      "probabilities": [
+        {
+          "option": "Visit the free art gallery",
+          "probability": 1
+        },
+        {
+          "option": "Take a long hike",
+          "probability": 3.0575177806246856e-11
+        },
+        {
+          "option": "Book a rooftop lunch",
+          "probability": 1.0092975066822163e-11
+        }
+      ],
+      "elapsed_seconds": 1.5341426910017617,
+      "input_tokens": 391
+    }
+  ],
+  "checks": [
+    "Real browser scoring and normalized distributions for every configured model",
+    "Input edits visibly mark old results as stale",
+    "Duplicate options rejected without scoring",
+    "Copy JSON returns the actual selected model result",
+    "Actual tokenizer rejects oversized input without truncation",
+    "Mobile layout has no horizontal overflow and real results remain usable",
+    "No browser script or Content Security Policy errors"
+  ],
+  "status": "passed",
+  "finished_utc": "2026-09-17T03:35:26.741Z"
+}

source/results/20260917-playground/concurrent-training.json

@@ -0,0 +1,343 @@
+{
+  "gui_run": "/home/andy/ai/opensysone/runs/20260917T032753Z-playground",
+  "gui_pid": 1467162,
+  "gui_source_commit": "e94e58ee311ab10ebe26fc46d4a5dde2280679f6",
+  "trainer_campaign": "/home/andy/ai/opensysone/runs/20260916T193741Z-24h",
+  "purpose": "Read-only coexistence observation after browser tests switched three real frozen models",
+  "observation_method": "Two process/memory/status samples about45seconds apart; GET /api/status only; no scoring/model requests from observer",
+  "launch_preflight_mem_available_bytes": 106415841280,
+  "samples": [
+    {
+      "observed_utc": "2026-09-17T03:35:55.926185+00:00",
+      "free_b": {
+        "exit_code": 0,
+        "stdout": "total        used        free      shared  buff/cache   available\nMem:     130594160640 43303702528 29233373184   164196352 36601188352 87290458112\nSwap:              0           0           0"
+      },
+      "gpu_processes": {
+        "exit_code": 0,
+        "stdout": "4674, /home/andy/ai/apps/llama.cpp/build/bin/llama-server, 170 MiB\n1096505, /home/andy/ai/envs/opensysone/bin/python, 16315 MiB\n1467162, /home/andy/ai/envs/opensysone/bin/python, 15655 MiB"
+      },
+      "gpu_status": {
+        "exit_code": 0,
+        "stdout": "71, 93 %, 44.89 W"
+      },
+      "processes": [
+        {
+          "role": "gui",
+          "pid": 1467162,
+          "command_matches": true,
+          "oom_score_adj": "0",
+          "status": {
+            "State": "S (sleeping)",
+            "VmHWM": "32281892 kB",
+            "VmRSS": "24261080 kB",
+            "Threads": "56"
+          }
+        },
+        {
+          "role": "trainer",
+          "pid": 1096505,
+          "command_matches": true,
+          "oom_score_adj": "0",
+          "status": {
+            "State": "S (sleeping)",
+            "VmHWM": "24653352 kB",
+            "VmRSS": "3402476 kB",
+            "Threads": "57"
+          }
+        }
+      ],
+      "gui_status": {
+        "http_status": 200,
+        "body": {
+          "status": "ready",
+          "loaded_model_id": "gx10-4b"
+        }
+      },
+      "gui_error_marker_counts": {
+        "model_load_failed": 0,
+        "model_score_failed": 0,
+        "Traceback": 0,
+        "OutOfMemoryError": 0
+      },
+      "training": {
+        "last_step": 3063,
+        "last_update": {
+          "step": 3063,
+          "loss": 0.014626714604673907,
+          "gradient_norm": 0.30683499574661255,
+          "seconds": 8.4302611759922,
+          "decisions": 4,
+          "branches": 12,
+          "actual_branch_tokens": 1218,
+          "padded_branch_tokens": 1218,
+          "peak_cuda_allocated_bytes": 16776178176,
+          "lr_factor": 0.977023980559572
+        },
+        "recent100_finite": true,
+        "recent100_median_step_seconds": 9.043510261006304,
+        "recent100_mean_step_seconds": 9.17212467954203,
+        "recent100_max_step_seconds": 11.7679517439974,
+        "recent4_updates": [
+          {
+            "step": 3060,
+            "loss": 0.019744501798413694,
+            "gradient_norm": 0.9511153101921082,
+            "seconds": 11.7679517439974,
+            "decisions": 4,
+            "branches": 13,
+            "actual_branch_tokens": 1523,
+            "padded_branch_tokens": 1523,
+            "peak_cuda_allocated_bytes": 16776178176,
+            "lr_factor": 0.9770701241265155
+          },
+          {
+            "step": 3061,
+            "loss": 0.3342024694720731,
+            "gradient_norm": 11.301324844360352,
+            "seconds": 11.187965510995127,
+            "decisions": 4,
+            "branches": 11,
+            "actual_branch_tokens": 1641,
+            "padded_branch_tokens": 1641,
+            "peak_cuda_allocated_bytes": 16776178176,
+            "lr_factor": 0.977054747970101
+          },
+          {
+            "step": 3062,
+            "loss": 0.02495433943113312,
+            "gradient_norm": 0.9082462191581726,
+            "seconds": 9.814839319995372,
+            "decisions": 4,
+            "branches": 14,
+            "actual_branch_tokens": 1375,
+            "padded_branch_tokens": 1375,
+            "peak_cuda_allocated_bytes": 16776178176,
+            "lr_factor": 0.9770393667810657
+          },
+          {
+            "step": 3063,
+            "loss": 0.014626714604673907,
+            "gradient_norm": 0.30683499574661255,
+            "seconds": 8.4302611759922,
+            "decisions": 4,
+            "branches": 12,
+            "actual_branch_tokens": 1218,
+            "padded_branch_tokens": 1218,
+            "peak_cuda_allocated_bytes": 16776178176,
+            "lr_factor": 0.977023980559572
+          }
+        ],
+        "peak_cuda_allocated_bytes": 16776178176,
+        "checkpoint_age_seconds": 585.4080212116241,
+        "heartbeat_utc": "2026-09-17T03:35:52.527546+00:00"
+      }
+    },
+    {
+      "observed_utc": "2026-09-17T03:36:41.021859+00:00",
+      "free_b": {
+        "exit_code": 0,
+        "stdout": "total        used        free      shared  buff/cache   available\nMem:     130594160640 43292459008 29244313600   164196352 36601479168 87301701632\nSwap:              0           0           0"
+      },
+      "gpu_processes": {
+        "exit_code": 0,
+        "stdout": "4674, /home/andy/ai/apps/llama.cpp/build/bin/llama-server, 170 MiB\n1096505, /home/andy/ai/envs/opensysone/bin/python, 16315 MiB\n1467162, /home/andy/ai/envs/opensysone/bin/python, 15655 MiB"
+      },
+      "gpu_status": {
+        "exit_code": 0,
+        "stdout": "71, 93 %, 45.40 W"
+      },
+      "processes": [
+        {
+          "role": "gui",
+          "pid": 1467162,
+          "command_matches": true,
+          "oom_score_adj": "0",
+          "status": {
+            "State": "S (sleeping)",
+            "VmHWM": "32281892 kB",
+            "VmRSS": "24261080 kB",
+            "Threads": "56"
+          }
+        },
+        {
+          "role": "trainer",
+          "pid": 1096505,
+          "command_matches": true,
+          "oom_score_adj": "0",
+          "status": {
+            "State": "S (sleeping)",
+            "VmHWM": "24653352 kB",
+            "VmRSS": "3402596 kB",
+            "Threads": "57"
+          }
+        }
+      ],
+      "gui_status": {
+        "http_status": 200,
+        "body": {
+          "status": "ready",
+          "loaded_model_id": "gx10-4b"
+        }
+      },
+      "gui_error_marker_counts": {
+        "model_load_failed": 0,
+        "model_score_failed": 0,
+        "Traceback": 0,
+        "OutOfMemoryError": 0
+      },
+      "training": {
+        "last_step": 3068,
+        "last_update": {
+          "step": 3068,
+          "loss": 0.001936116958859202,
+          "gradient_norm": 0.10404492169618607,
+          "seconds": 8.810512408002978,
+          "decisions": 4,
+          "branches": 12,
+          "actual_branch_tokens": 1407,
+          "padded_branch_tokens": 1407,
+          "peak_cuda_allocated_bytes": 16776178176,
+          "lr_factor": 0.9769469739709091
+        },
+        "recent100_finite": true,
+        "recent100_median_step_seconds": 9.03708774750703,
+        "recent100_mean_step_seconds": 9.169983857631742,
+        "recent100_max_step_seconds": 11.7679517439974,
+        "recent4_updates": [
+          {
+            "step": 3065,
+            "loss": 0.15298267499747453,
+            "gradient_norm": 6.990951061248779,
+            "seconds": 8.907861668005353,
+            "decisions": 4,
+            "branches": 10,
+            "actual_branch_tokens": 1726,
+            "padded_branch_tokens": 1726,
+            "peak_cuda_allocated_bytes": 16776178176,
+            "lr_factor": 0.9769931930198583
+          },
+          {
+            "step": 3066,
+            "loss": 0.0036661182439274853,
+            "gradient_norm": 0.1209888607263565,
+            "seconds": 8.316543419990921,
+            "decisions": 4,
+            "branches": 11,
+            "actual_branch_tokens": 1353,
+            "padded_branch_tokens": 1353,
+            "peak_cuda_allocated_bytes": 16776178176,
+            "lr_factor": 0.9769777917019633
+          },
+          {
+            "step": 3067,
+            "loss": 0.1977057716858326,
+            "gradient_norm": 11.540794372558594,
+            "seconds": 9.568666854989715,
+            "decisions": 4,
+            "branches": 11,
+            "actual_branch_tokens": 1853,
+            "padded_branch_tokens": 1853,
+            "peak_cuda_allocated_bytes": 16776178176,
+            "lr_factor": 0.9769623853522593
+          },
+          {
+            "step": 3068,
+            "loss": 0.001936116958859202,
+            "gradient_norm": 0.10404492169618607,
+            "seconds": 8.810512408002978,
+            "decisions": 4,
+            "branches": 12,
+            "actual_branch_tokens": 1407,
+            "padded_branch_tokens": 1407,
+            "peak_cuda_allocated_bytes": 16776178176,
+            "lr_factor": 0.9769469739709091
+          }
+        ],
+        "peak_cuda_allocated_bytes": 16776178176,
+        "checkpoint_age_seconds": 630.5036952495575,
+        "heartbeat_utc": "2026-09-17T03:36:37.538019+00:00"
+      }
+    }
+  ],
+  "observation_window": {
+    "elapsed_seconds": 45.095674,
+    "optimizer_updates_completed": 5,
+    "steps_per_minute": 6.6525228118333475,
+    "new_update_rows": [
+      {
+        "step": 3064,
+        "loss": 0.008182858524378389,
+        "gradient_norm": 0.19264547526836395,
+        "seconds": 9.048552577005466,
+        "decisions": 4,
+        "branches": 13,
+        "actual_branch_tokens": 1311,
+        "padded_branch_tokens": 1311,
+        "peak_cuda_allocated_bytes": 16776178176,
+        "lr_factor": 0.977008589305782
+      },
+      {
+        "step": 3065,
+        "loss": 0.15298267499747453,
+        "gradient_norm": 6.990951061248779,
+        "seconds": 8.907861668005353,
+        "decisions": 4,
+        "branches": 10,
+        "actual_branch_tokens": 1726,
+        "padded_branch_tokens": 1726,
+        "peak_cuda_allocated_bytes": 16776178176,
+        "lr_factor": 0.9769931930198583
+      },
+      {
+        "step": 3066,
+        "loss": 0.0036661182439274853,
+        "gradient_norm": 0.1209888607263565,
+        "seconds": 8.316543419990921,
+        "decisions": 4,
+        "branches": 11,
+        "actual_branch_tokens": 1353,
+        "padded_branch_tokens": 1353,
+        "peak_cuda_allocated_bytes": 16776178176,
+        "lr_factor": 0.9769777917019633
+      },
+      {
+        "step": 3067,
+        "loss": 0.1977057716858326,
+        "gradient_norm": 11.540794372558594,
+        "seconds": 9.568666854989715,
+        "decisions": 4,
+        "branches": 11,
+        "actual_branch_tokens": 1853,
+        "padded_branch_tokens": 1853,
+        "peak_cuda_allocated_bytes": 16776178176,
+        "lr_factor": 0.9769623853522593
+      },
+      {
+        "step": 3068,
+        "loss": 0.001936116958859202,
+        "gradient_norm": 0.10404492169618607,
+        "seconds": 8.810512408002978,
+        "decisions": 4,
+        "branches": 12,
+        "actual_branch_tokens": 1407,
+        "padded_branch_tokens": 1407,
+        "peak_cuda_allocated_bytes": 16776178176,
+        "lr_factor": 0.9769469739709091
+      }
+    ],
+    "new_loss_and_gradients_all_finite": true,
+    "mean_update_seconds": 8.930427385598886,
+    "median_update_seconds": 8.907861668005353,
+    "max_update_seconds": 9.568666854989715
+  },
+  "limitations": [
+    "No controlled baseline; recent100-step durations may include GUI activity.",
+    "GPU process memory from nvidia-smi includes context/cache overhead and is not the PyTorch allocated-memory cap.",
+    "Short observation verifies continued finite training and measured coexistence, not zero slowdown or long-term safety."
+  ],
+  "observer_loaded_models": false,
+  "observer_changed_processes": false,
+  "observer_accessed_credentials": false,
+  "reserved_predictions_accessed": false
+}

source/results/20260917-playground/frontend-fixture-check.json

@@ -0,0 +1,31 @@
+{
+  "scope": "Browser interaction tests with stub model responses; not model evidence",
+  "report": {
+    "fixture_only": true,
+    "checked_utc": "2026-09-17T03:33:23.412Z",
+    "checks": [
+      "catalog, initial example, empty state",
+      "keyboard submit, loading, genuine response rendering, ranked labels, CSP bar width",
+      "copy exact results JSON",
+      "edited inputs and changed model mark prior results stale without relabeling them",
+      "clear, empty input validation, reset",
+      "duplicate validation, all three examples",
+      "option text rendered safely without HTML execution",
+      "503 error, prior results preserved and marked stale, ready recovery",
+      "malformed probability rejection",
+      "390px mobile layout without horizontal overflow",
+      "320px layout without horizontal overflow",
+      "no localStorage persistence",
+      "model-list failure and retry",
+      "empty model catalog does not fabricate availability"
+    ],
+    "page_errors": [],
+    "csp_errors": [],
+    "screenshots": [
+      "desktop-empty.png",
+      "desktop-fixture-results.png",
+      "mobile-empty.png",
+      "mobile-fixture-results.png"
+    ]
+  }
+}

source/results/20260917-playground/launch.json

@@ -0,0 +1,38 @@
+{
+  "started_utc": "2026-09-17T03:34:51.304928+00:00",
+  "pid": 1467162,
+  "command": [
+    "/home/andy/ai/envs/opensysone/bin/python",
+    "/home/andy/projects/opensysone/playground.py",
+    "--models",
+    "/home/andy/ai/opensysone/runs/20260917T032753Z-playground/models.json",
+    "--port",
+    "18082",
+    "--device",
+    "cuda",
+    "--max-tokens",
+    "1024"
+  ],
+  "source_commit": "e94e58ee311ab10ebe26fc46d4a5dde2280679f6",
+  "source_sha256": {
+    "playground.py": "0eed2ab323dc73c7970b6d11fe41ddcb25e03f376af7e27891fd7e6c6a1e6e12",
+    "jev_harness.py": "4d4e979cb7ae352bcdacaaa6d64045e6b5e550b1a9721d4bad545045bee6c67f",
+    "experiment.py": "c036cbe9fbf6f71c1bb141dcfbecfb06a2f4bd697b070d2274b1931480e92031",
+    "training_model.py": "d5b0aefeeb5290816bc0b669aa0a8cbbe27f6a12b9cb23c141ac9b9ae9ee4e65",
+    "web/playground.html": "69495feec84f29a6efd06c7e45c1ddac3d37717a16b7b9036093609dfd24b4df",
+    "web/playground.css": "09969ed5332298048595879c930d1065850db3c915688ea2da343845e879f27b",
+    "web/playground.js": "b2ccc2dab0fb670fb3cd11e43f641cdab7c3752724351e7641f4dd3a4ca86646"
+  },
+  "model_catalog": "/home/andy/ai/opensysone/runs/20260917T032753Z-playground/models.json",
+  "model_catalog_sha256": "7a4e33a07801ab5fe918bb5a94c032900bd798513f6ab99608c9f06bb8c0572c",
+  "url": "http://127.0.0.1:18082",
+  "log": "/home/andy/ai/opensysone/runs/20260917T032753Z-playground/server.log",
+  "preflight": {
+    "free_b": "               total        used        free      shared  buff/cache   available\nMem:     130594160640 24178778112 70572675072   113401856 36546781184 106415382528\nSwap:              0           0           0\n",
+    "gpu_processes": "pid, process_name, used_gpu_memory [MiB]\n4674, /home/andy/ai/apps/llama.cpp/build/bin/llama-server, 170 MiB\n1096505, /home/andy/ai/envs/opensysone/bin/python, 16315 MiB\n",
+    "mem_available_bytes": 106415841280
+  },
+  "oom_score_adj": 0,
+  "inherited_secret_variables_removed": true,
+  "exit_status": "pending"
+}

source/results/20260917-playground/models.json

@@ -0,0 +1,32 @@
+{
+  "default_model": "gx10-4b",
+  "models": [
+    {
+      "id": "gx10-4b",
+      "label": "Qwen3 4B \u00b7 Main",
+      "description": "Main training run; selected checkpoint at step 2,500.",
+      "checkpoint": "/home/andy/ai/opensysone/runs/20260917T032753Z-playground/models/gx10-4b/best.pt",
+      "checkpoint_step": 2500,
+      "calibrated": false,
+      "sha256": "8956eb6c0cfbb02124aeefd99c3b418c55f55fdb9a64260350622d98dbba1aec"
+    },
+    {
+      "id": "spark-a-4b",
+      "label": "Qwen3 4B \u00b7 Lower rate",
+      "description": "Lower learning rate; selected checkpoint at step 2,500.",
+      "checkpoint": "/home/andy/ai/opensysone/runs/20260917T032753Z-playground/models/spark-a-4b/best.pt",
+      "checkpoint_step": 2500,
+      "calibrated": false,
+      "sha256": "b7923de78e41b88c353d27a13c7c66a0d5700aa578512a1ba3e5e34803486fb7"
+    },
+    {
+      "id": "spark-b-2b",
+      "label": "Qwen3.5 2B",
+      "description": "Smaller model; selected checkpoint at step 2,000.",
+      "checkpoint": "/home/andy/ai/opensysone/runs/20260917T032753Z-playground/models/spark-b-2b/best.pt",
+      "checkpoint_step": 2000,
+      "calibrated": false,
+      "sha256": "be08efdc9121a64547b5c9331d5908eeb0978c4fbc962885fd34f9758ec89d11"
+    }
+  ]
+}

source/results/20260917-playground/snapshots.json

@@ -0,0 +1,30 @@
+{
+  "frozen_utc": "2026-09-17T03:27:53.905122+00:00",
+  "base_weights_copied": false,
+  "snapshots": [
+    {
+      "id": "gx10-4b",
+      "sha256": "8956eb6c0cfbb02124aeefd99c3b418c55f55fdb9a64260350622d98dbba1aec",
+      "source": "/home/andy/ai/opensysone/exports/20260917T022835Z-hf-backup/artifacts/gx10-4b/best.pt",
+      "frozen": "/home/andy/ai/opensysone/runs/20260917T032753Z-playground/models/gx10-4b/best.pt",
+      "selected_step": 2500,
+      "source_commit": "4a60423c39d70f8d50472ce4f4f7fa4a4bd9fce1"
+    },
+    {
+      "id": "spark-a-4b",
+      "sha256": "b7923de78e41b88c353d27a13c7c66a0d5700aa578512a1ba3e5e34803486fb7",
+      "source": "/home/andy/ai/opensysone/exports/20260917T022835Z-hf-backup/artifacts/spark-a-4b/best.pt",
+      "frozen": "/home/andy/ai/opensysone/runs/20260917T032753Z-playground/models/spark-a-4b/best.pt",
+      "selected_step": 2500,
+      "source_commit": "4a60423c39d70f8d50472ce4f4f7fa4a4bd9fce1"
+    },
+    {
+      "id": "spark-b-2b",
+      "sha256": "be08efdc9121a64547b5c9331d5908eeb0978c4fbc962885fd34f9758ec89d11",
+      "source": "/home/andy/ai/opensysone/exports/20260917T022835Z-hf-backup/artifacts/spark-b-2b/best.pt",
+      "frozen": "/home/andy/ai/opensysone/runs/20260917T032753Z-playground/models/spark-b-2b/best.pt",
+      "selected_step": 2000,
+      "source_commit": "4a60423c39d70f8d50472ce4f4f7fa4a4bd9fce1"
+    }
+  ]
+}

source/scripts/verify_playground.cjs

@@ -0,0 +1,106 @@
+/* Real-browser smoke test against an already-running playground; no fixtures. */
+const assert = require('node:assert/strict');
+const fs = require('node:fs/promises');
+const path = require('node:path');
+const { chromium } = require(process.env.PLAYWRIGHT_MODULE || 'playwright-core');
+
+async function main() {
+  const url = process.argv[2] || 'http://127.0.0.1:18082';
+  const output = process.argv[3];
+  if (!output || !process.env.CHROMIUM_PATH) {
+    throw new Error('Provide URL, a new output directory and CHROMIUM_PATH.');
+  }
+  await fs.mkdir(output, { recursive: false });
+  const browser = await chromium.launch({
+    executablePath: process.env.CHROMIUM_PATH,
+    headless: true,
+    args: ['--disable-gpu', '--no-sandbox'],
+  });
+  const report = { started_utc: new Date().toISOString(), url, real_model_responses: [], checks: [] };
+  try {
+    const context = await browser.newContext({ viewport: { width: 1440, height: 1000 } });
+    const page = await context.newPage();
+    const errors = [];
+    page.on('pageerror', error => errors.push(error.message));
+    page.on('console', message => {
+      if (message.type() === 'error' && /Content Security Policy|Refused to/.test(message.text())) {
+        errors.push(message.text());
+      }
+    });
+    await page.goto(url);
+    await page.waitForFunction(() => !document.getElementById('model').disabled);
+    const ids = await page.locator('#model option').evaluateAll(options => options.map(option => option.value));
+    assert(ids.length >= 2, 'Expected multiple real models');
+    assert.match(await page.locator('#calibration-badge').innerText(), /Uncalibrated/);
+    const score = async (model) => {
+      await page.selectOption('#model', model);
+      await page.waitForFunction(() => !document.getElementById('submit').disabled);
+      const responsePromise = page.waitForResponse(response =>
+        new URL(response.url()).pathname === '/api/score' && response.request().method() === 'POST',
+      { timeout: 300000 });
+      await page.locator('#submit').click();
+      const response = await responsePromise;
+      const body = await response.json();
+      assert.equal(response.status(), 200, JSON.stringify(body));
+      assert.equal(body.model.id, model);
+      assert.equal(body.probabilities.length, 3);
+      assert(Math.abs(body.probabilities.reduce((sum, row) => sum + row.probability, 0) - 1) < 1e-5);
+      assert(body.probabilities.every(row => Number.isFinite(row.probability) && row.probability >= 0));
+      await page.locator('#results-content').waitFor({ state: 'visible' });
+      assert.equal(await page.locator('.probability-item').count(), 3);
+      assert.equal(await page.locator('#stale-notice').isVisible(), false);
+      report.real_model_responses.push(body);
+      return body;
+    };
+    for (const [index, model] of ids.entries()) {
+      await score(model);
+      if (!index) await page.screenshot({ path: path.join(output, 'desktop.png'), fullPage: true });
+    }
+    report.checks.push('Real browser scoring and normalized distributions for every configured model');
+    await page.locator('#question').fill('Which choice fits the context best?');
+    assert.equal(await page.locator('#stale-notice').isVisible(), true);
+    report.checks.push('Input edits visibly mark old results as stale');
+    await page.locator('#options').fill('Duplicate\nDuplicate');
+    await page.locator('#submit').click();
+    assert.match(await page.locator('#error-message').innerText(), /duplicate|different/i);
+    report.checks.push('Duplicate options rejected without scoring');
+    await page.locator('#reset').click();
+    await score(ids[0]);
+    await context.grantPermissions(['clipboard-read', 'clipboard-write']);
+    await page.locator('#copy-results').click();
+    await page.waitForFunction(() => document.getElementById('copy-label').textContent === 'Copied');
+    assert.equal(await page.locator('#copy-label').innerText(), 'Copied');
+    const copied = JSON.parse(await page.evaluate(() => navigator.clipboard.readText()));
+    assert.equal(copied.model.id, ids[0]);
+    report.checks.push('Copy JSON returns the actual selected model result');
+    const tokenError = await page.evaluate(async () => {
+      const response = await fetch('/api/score', { method: 'POST', headers: { 'Content-Type': 'application/json' },
+        body: JSON.stringify({ model: document.getElementById('model').value,
+          text: 'A very long invented context. '.repeat(1500), question: 'Which option?', options: ['First', 'Second'] }) });
+      return { status: response.status, body: await response.json() };
+    });
+    assert.equal(tokenError.status, 422);
+    assert.equal(tokenError.body.code, 'token_limit');
+    report.checks.push('Actual tokenizer rejects oversized input without truncation');
+    await page.setViewportSize({ width: 390, height: 844 });
+    await page.locator('#reset').click();
+    await score(ids[0]);
+    assert(await page.evaluate(() => document.documentElement.scrollWidth <= window.innerWidth));
+    await page.screenshot({ path: path.join(output, 'mobile.png'), fullPage: true });
+    report.checks.push('Mobile layout has no horizontal overflow and real results remain usable');
+    assert.deepEqual(errors, []);
+    report.checks.push('No browser script or Content Security Policy errors');
+    report.status = 'passed';
+  } catch (error) {
+    report.status = 'failed';
+    report.error = error.message;
+    throw error;
+  } finally {
+    report.finished_utc = new Date().toISOString();
+    await fs.writeFile(path.join(output, 'browser-verification.json'), JSON.stringify(report, null, 2) + '\n');
+    await browser.close();
+  }
+  console.log(JSON.stringify({ status: report.status, models: report.real_model_responses.map(row => row.model.id), output }));
+}
+
+main().catch(error => { console.error(error.message); process.exitCode = 1; });

source/tests/test_playground.py

@@ -0,0 +1,196 @@
+import contextlib
+import gc
+import http.client
+import io
+import json
+from pathlib import Path
+import tempfile
+import threading
+import unittest
+import weakref
+
+import playground
+
+
+class PlaygroundTests(unittest.TestCase):
+    def setUp(self):
+        temporary = tempfile.TemporaryDirectory()
+        self.addCleanup(temporary.cleanup)
+        self.root = Path(temporary.name)
+        self.rows = []
+        for index, name in enumerate(('a', 'b')):
+            checkpoint = self.root / (name + '.pt')
+            checkpoint.write_bytes(('fake-' + name).encode())
+            self.rows.append({'id': name, 'label': name.upper(), 'description': 'Fixture model',
+                'checkpoint': str(checkpoint), 'checkpoint_step': index, 'calibrated': bool(index),
+                'sha256': playground.checksum(checkpoint)})
+        self.config = self.root / 'models.json'
+        self.config.write_text(json.dumps({'models': self.rows, 'default_model': 'a'}))
+        self.catalog, self.default = playground.load_catalog(self.config)
+        self.request = {'model': 'a', 'text': 'Context', 'question': 'Question?', 'options': [' Yes ', 'No']}
+
+    def service(self, blocking=None):
+        events, references, payloads = [], [], []
+        class Backend:
+            def __init__(self, checkpoint, device, max_tokens):
+                self.id = Path(checkpoint).stem
+                self.calibrated, self.max_tokens = self.id == 'b', max_tokens
+                events.append(('load', self.id, threading.get_ident()))
+                references.append(weakref.ref(self))
+                if blocking is not None:
+                    blocking[0].set()
+                    if not blocking[1].wait(5):
+                        raise RuntimeError('Test synchronization timed out')
+
+            def __call__(self, payload):
+                events.append(('score', self.id, threading.get_ident()))
+                payloads.append(payload)
+                if payload['state'] == 'too long':
+                    raise ValueError('Input has 1025 tokens; limit is 1024; no truncation')
+                if payload['state'] == 'private-error':
+                    raise RuntimeError('/private/checkpoint/secret.pt prompt text')
+                options = payload['questions']['answer']['criteria']
+                return {'answers': {'answer': {'type': 'choice', 'choice': next(iter(options)), 'confidence': 0.2, 'probabilities': {
+                    option: probability for option, probability in zip(options, (0.75, 0.25))}}},
+                    'usage': {'input_tokens': 42}}
+
+            def __del__(self):
+                events.append(('destroy', self.id, threading.get_ident()))
+
+        def cleanup(device):
+            gc.collect()
+            events.append(('cleanup', None, threading.get_ident()))
+        service = playground.ModelService(self.catalog, self.default, device='cpu', backend_factory=Backend,
+            preflight=lambda device: events.append(('preflight', None, threading.get_ident())), cleanup=cleanup)
+        self.addCleanup(service.close)
+        return service, events, references, payloads
+
+    def test_catalog_allowlist_frozen_hash_and_public_metadata(self):
+        service, _, _, _ = self.service()
+        public = service.models()
+        self.assertEqual(public['default_model'], 'a')
+        self.assertNotIn('checkpoint', public['models'][0])
+        self.assertNotIn('sha256', public['models'][0])
+        with self.assertRaises(playground.RequestError) as error:
+            service.score({**self.request, 'model': '/tmp/arbitrary.pt'})
+        self.assertEqual(error.exception.code, 'unknown_model')
+        (self.root / 'a.pt').write_bytes(b'changed after startup')
+        with contextlib.redirect_stdout(io.StringIO()), self.assertRaises(playground.RequestError) as error:
+            service.score(self.request)
+        self.assertEqual(error.exception.code, 'model_unavailable')
+        self.assertIsNone(service.backend)
+        self.assertEqual(service.status()['status'], 'error')
+
+    def test_validation_preserves_options_and_reports_token_limit(self):
+        service, _, _, payloads = self.service()
+        for changes in ({'options': ['x']}, {'options': ['x', 'x']}, {'options': ['x', ' ']},
+                        {'options': ['x', 1]}, {'text': ''}, {'question': None}, {'checkpoint': '/tmp/other'}):
+            with self.assertRaises(playground.RequestError):
+                service.score({**self.request, **changes})
+        result = service.score(self.request)
+        self.assertEqual(result['probabilities'], [{'option': ' Yes ', 'probability': 0.75},
+                                                    {'option': 'No', 'probability': 0.25}])
+        self.assertEqual(payloads[0]['questions']['answer']['criteria'], {' Yes ': None, 'No': None})
+        self.assertEqual(result['input_tokens'], 42)
+        self.assertFalse(result['model']['calibrated'])
+        with self.assertRaises(playground.RequestError) as error:
+            service.score({**self.request, 'text': 'too long'})
+        self.assertEqual((error.exception.status, error.exception.code), (422, 'token_limit'))
+        self.assertIn('1025', error.exception.message)
+        capture = io.StringIO()
+        with contextlib.redirect_stdout(capture), self.assertRaises(playground.RequestError) as error:
+            service.score({**self.request, 'text': 'private-error'})
+        self.assertNotIn('/private', error.exception.message + capture.getvalue())
+        self.assertNotIn('prompt text', capture.getvalue())
+        self.assertIn('RuntimeError', capture.getvalue())
+
+    def test_switch_frees_old_model_and_all_backend_work_uses_one_thread(self):
+        service, events, references, _ = self.service()
+        service.score(self.request)
+        service.score({**self.request, 'model': 'b'})
+        self.assertIsNone(references[0]())
+        self.assertIsNotNone(references[1]())
+        self.assertEqual(len({event[2] for event in events}), 1)
+        self.assertNotEqual(events[0][2], threading.get_ident())
+        destroy = next(index for index, event in enumerate(events) if event[:2] == ('destroy', 'a'))
+        load = next(index for index, event in enumerate(events) if event[:2] == ('load', 'b'))
+        self.assertLess(destroy, load)
+        self.assertTrue(any(event[0] == 'cleanup' for event in events[destroy + 1:load]))
+        self.assertEqual(service.status(), {'status': 'ready', 'loaded_model_id': 'b'})
+
+    def test_concurrent_score_is_busy_and_status_remains_responsive(self):
+        entered, release = threading.Event(), threading.Event()
+        service, _, _, _ = self.service(blocking=(entered, release))
+        results = []
+        worker = threading.Thread(target=lambda: results.append(service.score(self.request)))
+        worker.start()
+        try:
+            self.assertTrue(entered.wait(2))
+            self.assertEqual(service.status()['status'], 'loading')
+            with self.assertRaises(playground.RequestError) as error:
+                service.score(self.request)
+            self.assertEqual((error.exception.status, error.exception.code), (503, 'busy'))
+        finally:
+            release.set()
+            worker.join(5)
+        self.assertFalse(worker.is_alive())
+        self.assertEqual(len(results), 1)
+
+    def test_calibration_metadata_must_match_actual_backend(self):
+        self.catalog['a']['calibrated'] = True
+        service, _, references, _ = self.service()
+        with contextlib.redirect_stdout(io.StringIO()), self.assertRaises(playground.RequestError):
+            service.score(self.request)
+        self.assertIsNone(service.backend)
+        self.assertIsNone(references[0]())
+
+    def test_retained_scoring_error_cannot_hold_old_model_after_switch(self):
+        service, _, references, _ = self.service()
+        with contextlib.redirect_stdout(io.StringIO()), self.assertRaises(playground.RequestError) as caught:
+            service.score({**self.request, 'text': 'private-error'})
+        retained_error = caught.exception
+        service.score({**self.request, 'model': 'b'})
+        self.assertIsNotNone(retained_error)
+        self.assertIsNone(references[0]())
+
+    def test_http_assets_origin_limits_and_json_errors(self):
+        service, _, _, _ = self.service()
+        for name in ('playground.html', 'playground.css', 'playground.js'):
+            (self.root / name).write_text('fixture ' + name)
+        server = playground.create_server(service, port=0, web_root=self.root)
+        worker = threading.Thread(target=server.serve_forever, daemon=True)
+        worker.start()
+        self.addCleanup(worker.join, 2)
+        self.addCleanup(server.server_close)
+        self.addCleanup(server.shutdown)
+        def request(method, path, body=None, headers=None):
+            connection = http.client.HTTPConnection('127.0.0.1', server.server_port, timeout=3)
+            connection.request(method, path, body=body, headers=headers or {})
+            response = connection.getresponse()
+            content = response.read()
+            result = response.status, dict(response.getheaders()), content
+            connection.close()
+            return result
+        status, headers, content = request('GET', '/')
+        self.assertEqual(status, 200)
+        self.assertEqual(headers['X-Content-Type-Options'], 'nosniff')
+        self.assertIn("frame-ancestors 'none'", headers['Content-Security-Policy'])
+        self.assertIn(b'playground.html', content)
+        self.assertEqual(request('GET', '/../../models.json')[0], 404)
+        self.assertEqual(request('GET', '/', headers={'Host': 'attacker.example'})[0], 403)
+        payload = json.dumps(self.request)
+        matching = {'Host': 'localhost:9999', 'Origin': 'http://localhost:9999', 'Content-Type': 'application/json'}
+        self.assertEqual(request('POST', '/api/score', payload, matching)[0], 200)
+        self.assertEqual(request('POST', '/api/score', payload, {**matching, 'Origin': 'http://evil.example'})[0], 403)
+        self.assertEqual(request('POST', '/api/score', payload, {'Content-Type': 'text/plain'})[0], 415)
+        self.assertEqual(request('POST', '/api/score', '{', {'Content-Type': 'application/json'})[0], 400)
+        self.assertEqual(request('POST', '/api/score', b'', {'Content-Type': 'application/json',
+                         'Content-Length': str(playground.MAX_BODY_BYTES + 1)})[0], 413)
+        status, headers, content = request('PUT', '/api/score')
+        self.assertEqual(status, 501)
+        self.assertIn('application/json', headers['Content-Type'])
+        self.assertIn('error', json.loads(content))
+
+
+if __name__ == '__main__':
+    unittest.main()

source/web/playground.css

@@ -0,0 +1,187 @@
+:root {
+  color-scheme: light;
+  --paper: #f5f3ed;
+  --surface: #fffefa;
+  --ink: #242c27;
+  --muted: #606a60;
+  --line: #dedfd5;
+  --green: #285c46;
+  --green-dark: #1d4634;
+  --sage: #edf0e7;
+  --error: #903f2e;
+  --font: "Segoe UI", -apple-system, BlinkMacSystemFont, "Helvetica Neue", sans-serif;
+  font-family: var(--font);
+  font-synthesis: none;
+  -webkit-font-smoothing: antialiased;
+  color: var(--ink);
+  background: var(--paper);
+}
+
+* { box-sizing: border-box; }
+body { margin: 0; }
+button, input, textarea, select { font: inherit; }
+button, select { cursor: pointer; }
+button { -webkit-tap-highlight-color: transparent; }
+button:disabled, select:disabled { cursor: default; }
+button:disabled { opacity: .48; }
+button, a, textarea, select { outline-offset: 4px; }
+:focus-visible { outline: 2px solid var(--green); }
+button, a { touch-action: manipulation; }
+[hidden] { display: none !important; }
+svg { flex-shrink: 0; }
+a { color: inherit; }
+.page-shell { width: min(1260px, calc(100% - 96px)); margin: 0 auto; }
+.site-header { height: 100px; display: flex; align-items: center; justify-content: space-between; border-bottom: 1px solid var(--line); }
+.wordmark { display: inline-flex; align-items: center; gap: 12px; text-decoration: none; font-size: 19px; font-weight: 650; letter-spacing: -.7px; }
+.brand-mark { width: 27px; height: 27px; color: var(--green); }
+.connection { display: flex; align-items: center; gap: 8px; font-size: 12px; color: var(--muted); }
+.status-dot { width: 6px; height: 6px; border-radius: 50%; background: #a4a79b; }
+.connection[data-state="ready"] .status-dot { background: var(--green); box-shadow: 0 0 0 4px #285c4610; }
+.connection[data-state="busy"] .status-dot { background: #ae7c2c; animation: breathe 1.8s ease-in-out infinite; }
+.connection[data-state="error"] .status-dot { background: var(--error); }
+.intro { display: grid; grid-template-columns: 1.3fr 1fr; align-items: center; gap: 72px; padding: 56px 0 48px; }
+.eyebrow { margin: 0 0 18px; font-size: 11px; font-weight: 650; text-transform: uppercase; letter-spacing: 2px; color: var(--green); }
+h1 { margin: 0; font-family: Georgia, "Times New Roman", serif; font-size: clamp(38px, 4.3vw, 57px); line-height: 1.08; font-weight: 400; letter-spacing: -2px; }
+.intro-description { margin: 21px 0 0; color: var(--muted); font-size: 15px; line-height: 1.65; }
+.model-panel { padding: 4px 0 0 27px; border-left: 1px solid var(--line); }
+.field-label { display: block; margin-bottom: 10px; font-size: 13px; font-weight: 650; letter-spacing: -.1px; }
+.select-wrap { position: relative; }
+.select-wrap svg { pointer-events: none; position: absolute; right: 15px; top: 17px; width: 18px; height: 18px; color: var(--green); }
+select { appearance: none; -webkit-appearance: none; width: 100%; min-height: 52px; padding: 13px 41px 13px 15px; border: 1px solid #cdd3c7; border-radius: 8px; background: var(--surface); color: var(--ink); font-weight: 550; font-size: 14px; text-overflow: ellipsis; }
+select:disabled { color: var(--muted); }
+.model-description { min-height: 36px; margin: 11px 0 12px; font-size: 12px; line-height: 1.55; color: var(--muted); }
+.model-details { display: flex; gap: 10px 12px; align-items: center; flex-wrap: wrap; min-height: 25px; }
+.calibration-badge { color: #775b30; background: #eee6d6; border: 1px solid #e2d8c5; border-radius: 4px; padding: 4px 7px; font-size: 10px; line-height: 1.2; font-weight: 600; }
+.calibration-badge[data-calibrated="true"] { background: #e4ecdf; color: var(--green); border-color: #d3e0cb; }
+.model-limit { font-size: 11px; color: var(--muted); }
+.retry-button { margin-top: 12px; }
+.workspace { display: grid; grid-template-columns: 1.06fr 1fr; border: 1px solid #d6dbce; border-radius: 14px; overflow: hidden; background: var(--surface); box-shadow: 0 6px 32px #283b2710; }
+.editor-panel, .results-panel { min-width: 0; padding: 30px; }
+.results-panel { border-left: 1px solid #d6dbce; background: var(--sage); display: flex; flex-direction: column; }
+.panel-heading { display: flex; align-items: center; justify-content: space-between; gap: 16px; min-height: 26px; margin-bottom: 25px; }
+h2 { margin: 0; font-size: 16px; font-weight: 650; letter-spacing: -.35px; }
+.text-button { border: 0; padding: 3px 0; background: transparent; color: var(--green); font-size: 12px; line-height: 1.6; text-decoration-thickness: 1px; text-underline-offset: 3px; }
+.text-button:hover:not(:disabled) { text-decoration: underline; }
+.examples { margin-bottom: 27px; }
+.example-label { display: block; font-size: 11px; color: var(--muted); margin-bottom: 10px; }
+.example-buttons { display: flex; flex-wrap: wrap; gap: 7px; }
+.example-button { padding: 7px 10px; border: 1px solid #dce0d5; border-radius: 5px; background: transparent; color: #5b655b; font-size: 11px; line-height: 1.35; transition: background .15s, border-color .15s; }
+.example-button:hover:not(:disabled) { background: #f0f2eb; border-color: #aebca7; }
+.example-button[aria-pressed="true"] { color: var(--green-dark); border-color: #a9bda0; background: #e9efdf; }
+.field-group { margin-bottom: 22px; }
+.label-row { display: flex; justify-content: space-between; align-items: baseline; gap: 12px; }
+.label-row .field-hint { margin: 0; font-size: 10px; flex-shrink: 0; }
+textarea { display: block; width: 100%; border: 1px solid #dce0d5; border-radius: 7px; padding: 13px 14px; color: #384238; background: #fbfcf7; line-height: 1.6; font-size: 13px; resize: vertical; min-height: 80px; transition: border-color .15s, background .15s; }
+textarea:hover { border-color: #bac5b3; }
+textarea:focus { border-color: var(--green); background: #fffefa; }
+textarea[readonly] { color: #687166; background: #f3f5ee; }
+textarea[aria-invalid="true"] { border-color: var(--error); }
+textarea::placeholder { color: #8b9289; }
+#context { min-height: 185px; }
+.question-input { min-height: 71px; }
+.options-input { min-height: 118px; line-height: 1.85; }
+.field-hint { margin: 8px 0 0; font-size: 11px; line-height: 1.55; color: var(--muted); }
+.options-group { margin-bottom: 25px; }
+.error-message { margin: -3px 0 18px; border: 1px solid #e5c9bd; border-radius: 6px; padding: 12px 13px; color: var(--error); background: #fcf0e9; font-size: 12px; line-height: 1.55; overflow-wrap: anywhere; }
+.form-actions { display: flex; gap: 20px; align-items: center; }
+.primary-button { display: inline-flex; justify-content: center; align-items: center; gap: 17px; min-height: 48px; border: 1px solid var(--green); border-radius: 6px; padding: 12px 19px; background: var(--green); color: #fffefa; font-size: 13px; font-weight: 550; box-shadow: 0 2px 0 #1a392716; transition: background .15s, transform .15s; }
+.primary-button:hover:not(:disabled) { background: var(--green-dark); transform: translateY(-1px); }
+.primary-button:active:not(:disabled) { transform: translateY(0); }
+.primary-button:disabled { opacity: .56; }
+.primary-button svg { width: 17px; height: 17px; }
+.button-spinner { display: none; width: 14px; height: 14px; border: 1.5px solid #ffffff55; border-top-color: white; border-radius: 50%; animation: spin .8s linear infinite; }
+.primary-button[data-loading="true"] .button-spinner { display: block; }
+.primary-button[data-loading="true"] .submit-arrow { display: none; }
+.keyboard-hint { color: #687360; font-size: 10px; white-space: nowrap; }
+kbd { font-family: inherit; }
+.editor-footnote { margin-top: 24px; padding-top: 17px; border-top: 1px solid #e8eade; display: flex; flex-wrap: wrap; align-items: center; justify-content: space-between; gap: 10px; font-size: 10px; color: #687360; line-height: 1.5; }
+.editor-footnote .text-button { font-size: 10px; }
+.results-count { color: #617053; font-size: 11px; }
+.empty-state, .loading-state { margin: auto 0; padding: 65px 15px 100px; text-align: center; }
+.empty-state h3, .loading-state h3 { margin: 25px 0 12px; font-family: Georgia, "Times New Roman", serif; font-size: 24px; font-weight: 400; letter-spacing: -.65px; }
+.empty-state p, .loading-state p { margin: 0; color: #64705d; font-size: 12px; line-height: 1.8; }
+.empty-mark { display: flex; align-items: center; justify-content: center; gap: 7px; width: 83px; height: 83px; margin: 0 auto; border: 1px solid #d3ddc8; border-radius: 50%; background: #e6ecdd; }
+.empty-mark span { width: 6px; height: 6px; background: #9bac8b; border-radius: 50%; }
+.empty-mark span:nth-child(2) { width: 8px; height: 8px; background: #6c885c; }
+.loading-orbit { width: 54px; height: 54px; margin: 0 auto; border-radius: 50%; border: 1px solid #c8d5bd; border-top: 2px solid var(--green); animation: spin 1.3s linear infinite; }
+.elapsed-timer { display: inline-block; margin-top: 21px; color: var(--green); font-size: 12px; font-variant-numeric: tabular-nums; }
+.results-content { padding-top: 3px; }
+.result-question { font-family: Georgia, "Times New Roman", serif; font-size: 24px; line-height: 1.35; letter-spacing: -.55px; margin: 0 0 32px; overflow-wrap: anywhere; }
+.stale-notice { margin: -14px 0 25px; border-left: 2px solid #a8b897; padding-left: 10px; font-size: 11px; color: #69795d; line-height: 1.5; }
+.probability-list { margin: 0; padding: 0; list-style: none; }
+.probability-item { margin-bottom: 25px; }
+.probability-row { display: grid; grid-template-columns: 22px minmax(0, 1fr) auto; align-items: baseline; column-gap: 9px; margin-bottom: 10px; }
+.option-index { font-size: 9px; font-weight: 600; color: #65735a; letter-spacing: .6px; }
+.option-text { color: #54614c; font-size: 13px; line-height: 1.6; overflow-wrap: anywhere; white-space: pre-wrap; }
+.option-probability { font-size: 17px; letter-spacing: -.4px; font-weight: 550; color: #607151; font-variant-numeric: tabular-nums; }
+.probability-item:first-child .option-text { color: var(--green-dark); font-weight: 600; }
+.probability-item:first-child .option-probability { color: var(--green-dark); font-size: 23px; font-weight: 600; }
+.probability-track { margin-left: 31px; height: 9px; background: #dee5d5; border-radius: 2px; overflow: hidden; }
+.probability-fill { height: 100%; background: #a5b997; border-radius: 2px; transition: width .45s ease-out; }
+.probability-item:first-child .probability-fill { background: var(--green); }
+.probability-note { font-size: 10px; line-height: 1.7; color: #65735a; margin: 33px 0 28px; }
+.result-footer { padding-top: 20px; border-top: 1px solid #d6dfcd; display: flex; align-items: center; justify-content: space-between; gap: 15px; }
+.result-metadata { display: flex; flex-direction: column; gap: 6px; font-size: 10px; color: #606e56; line-height: 1.5; min-width: 0; }
+.result-model { color: #52634a; font-weight: 550; overflow-wrap: anywhere; }
+.copy-button { flex-shrink: 0; display: flex; align-items: center; gap: 5px; padding: 7px 9px; background: transparent; border: 1px solid #cbd7be; border-radius: 5px; color: #596d4d; font-size: 10px; }
+.copy-button:hover { background: #e2ead9; }
+.copy-button svg { width: 15px; height: 15px; }
+.site-footer { display: flex; justify-content: space-between; gap: 20px; padding: 24px 0 36px; font-size: 10px; color: #687360; letter-spacing: .1px; }
+.site-footer span:first-child { font-weight: 600; }
+.skip-link { position: fixed; z-index: 10; left: 15px; top: -80px; padding: 12px 16px; color: white; background: var(--green); border-radius: 5px; }
+.skip-link:not(:focus) { clip-path: inset(50%); clip: rect(0, 0, 0, 0); width: 1px; height: 1px; padding: 0; overflow: hidden; white-space: nowrap; }
+.skip-link:focus { top: 15px; }
+.sr-only { position: absolute; width: 1px; height: 1px; padding: 0; margin: -1px; overflow: hidden; clip: rect(0,0,0,0); white-space: nowrap; border: 0; }
+.noscript-message { margin: 25px auto; max-width: 600px; text-align: center; color: var(--error); }
+
+@keyframes spin { to { transform: rotate(360deg); } }
+@keyframes breathe { 50% { opacity: .35; } }
+@media (min-width: 1450px) { .intro { padding-top: 64px; padding-bottom: 54px; } }
+@media (max-width: 1000px) {
+  .page-shell { width: calc(100% - 56px); }
+  .intro { gap: 35px; grid-template-columns: 1.1fr 1fr; }
+  .editor-panel, .results-panel { padding: 24px; }
+  .keyboard-hint { display: none; }
+  .model-panel { padding-left: 23px; }
+  .result-footer { flex-wrap: wrap; }
+}
+@media (max-width: 760px) {
+  .page-shell { width: calc(100% - 36px); max-width: 600px; }
+  .site-header { height: 77px; }
+  .wordmark { font-size: 17px; }
+  .intro { display: block; padding: 35px 0 29px; }
+  h1 { font-size: 43px; letter-spacing: -1.7px; }
+  .eyebrow { font-size: 10px; margin-bottom: 15px; }
+  .intro-description { font-size: 14px; margin-top: 17px; }
+  .model-panel { margin-top: 26px; padding: 0; border: 0; }
+  .model-description { min-height: 0; }
+  .workspace { grid-template-columns: 1fr; border-radius: 10px; }
+  .editor-panel, .results-panel { padding: 23px 20px; }
+  .results-panel { border-left: 0; border-top: 1px solid #d6dbce; min-height: 340px; }
+  .empty-state, .loading-state { padding: 36px 10px 46px; }
+  .empty-mark { height: 65px; width: 65px; }
+  .empty-state h3, .loading-state h3 { margin-top: 21px; }
+  .form-actions { display: block; }
+  .primary-button { width: 100%; min-height: 49px; }
+  textarea { font-size: 16px; }
+  .question-input { min-height: 81px; }
+  .options-input { min-height: 137px; }
+  select { font-size: 16px; }
+  .text-button { min-height: 32px; }
+  .example-button { padding: 10px; }
+  .copy-button { min-height: 36px; padding: 8px 11px; }
+  .result-question { margin-top: 3px; font-size: 23px; }
+  .site-footer { padding-bottom: 25px; }
+}
+@media (max-width: 380px) {
+  .page-shell { width: calc(100% - 24px); }
+  h1 { font-size: 38px; }
+  .editor-panel, .results-panel { padding: 20px 16px; }
+  .connection { font-size: 10px; }
+  .example-button { font-size: 10px; padding: 9px 8px; }
+  .label-row .field-hint { font-size: 9px; }
+  .desktop-break { display: none; }
+}
+@media (prefers-reduced-motion: reduce) {
+  *, *::before, *::after { animation-duration: .01ms !important; animation-iteration-count: 1 !important; transition-duration: .01ms !important; scroll-behavior: auto !important; }
+}

source/web/playground.html

@@ -0,0 +1,147 @@
+<!doctype html>
+<html lang="en">
+<head>
+  <meta charset="utf-8">
+  <meta name="viewport" content="width=device-width, initial-scale=1">
+  <meta name="theme-color" content="#f5f3ed">
+  <meta name="description" content="Give OpenSysOne a context, a question and possible answers. Explore the probabilities it assigns to each choice.">
+  <title>Model playground · OpenSysOne</title>
+  <link rel="stylesheet" href="/playground.css">
+  <script src="/playground.js" defer></script>
+</head>
+<body>
+  <a class="skip-link" href="#context">Skip to the playground</a>
+  <div class="page-shell">
+    <header class="site-header">
+      <a class="wordmark" href="/" aria-label="OpenSysOne playground home">
+        <svg class="brand-mark" viewBox="0 0 28 28" fill="none" aria-hidden="true"><path d="M5 7h18M5 14h12M5 21h6" stroke="currentColor" stroke-width="4" stroke-linecap="round"/></svg>
+        <span>OpenSysOne</span>
+      </a>
+      <div class="connection" id="connection" data-state="connecting" role="status">
+        <span class="status-dot" aria-hidden="true"></span>
+        <span id="connection-text">Connecting</span>
+      </div>
+    </header>
+
+    <main>
+      <section class="intro" aria-labelledby="page-title">
+        <div class="intro-copy">
+          <p class="eyebrow">Model playground</p>
+          <h1 id="page-title">A little context.<br>A clearer choice.</h1>
+          <p class="intro-description">Give the model a question and possible answers.<br class="desktop-break"> See how it weighs each one.</p>
+        </div>
+        <div class="model-panel">
+          <label class="field-label" for="model">Choose a model</label>
+          <div class="select-wrap">
+            <select id="model" name="model" disabled aria-describedby="model-description model-limit">
+              <option value="">Loading available models…</option>
+            </select>
+            <svg viewBox="0 0 20 20" fill="none" aria-hidden="true"><path d="m6 8 4 4 4-4" stroke="currentColor" stroke-width="1.6" stroke-linecap="round" stroke-linejoin="round"/></svg>
+          </div>
+          <p class="model-description" id="model-description">Finding available checkpoints.</p>
+          <div class="model-details">
+            <span class="calibration-badge" id="calibration-badge" hidden></span>
+            <span id="model-limit" class="model-limit"></span>
+          </div>
+          <button class="text-button retry-button" id="retry-models" type="button" hidden>Retry connection <span aria-hidden="true">↗</span></button>
+        </div>
+      </section>
+
+      <div class="workspace">
+        <section class="editor-panel" aria-labelledby="editor-title">
+          <div class="panel-heading">
+            <h2 id="editor-title">Your decision</h2>
+            <button type="button" class="text-button" id="clear">Clear</button>
+          </div>
+          <div class="examples" aria-label="Try an example">
+            <span class="example-label">Try an example</span>
+            <div class="example-buttons">
+              <button type="button" class="example-button" data-example="0" aria-pressed="true">Weekend plans</button>
+              <button type="button" class="example-button" data-example="1" aria-pressed="false">Parcel handoff</button>
+              <button type="button" class="example-button" data-example="2" aria-pressed="false">Studio booking</button>
+            </div>
+          </div>
+
+          <form id="decision-form" novalidate>
+            <div class="field-group">
+              <div class="label-row">
+                <label class="field-label" for="context">Context</label>
+                <span class="field-hint" id="context-count">0 characters</span>
+              </div>
+              <textarea id="context" name="context" rows="7" placeholder="Add the information the model should use…" aria-describedby="context-hint context-count" required>Ari and Noor are planning Saturday afternoon. Rain is forecast from noon onward. They want an indoor activity, enjoy art, and have a small budget. The city gallery has a free exhibition open until 6 pm. The hiking trail has no shelter, and the rooftop restaurant is expensive.</textarea>
+              <p class="field-hint" id="context-hint">A situation, a message, a few facts — use your own words.</p>
+            </div>
+            <div class="field-group">
+              <label class="field-label" for="question">Question</label>
+              <textarea class="question-input" id="question" name="question" rows="2" placeholder="What would you like the model to decide?" required>Which plan best fits their preferences?</textarea>
+            </div>
+            <div class="field-group options-group">
+              <div class="label-row">
+                <label class="field-label" for="options">Possible answers</label>
+                <span class="field-hint" id="option-count">3 options</span>
+              </div>
+              <textarea class="options-input" id="options" name="options" rows="4" placeholder="First answer&#10;Second answer&#10;Another possibility" aria-describedby="options-hint option-count" required>Visit the free art gallery
+Take a long hike
+Book a rooftop lunch</textarea>
+              <p class="field-hint" id="options-hint">One answer per line. Include at least two distinct options.</p>
+            </div>
+            <div class="error-message" id="error-message" role="alert" hidden></div>
+            <div class="form-actions">
+              <button class="primary-button" id="submit" type="submit" disabled>
+                <span class="button-spinner" aria-hidden="true"></span>
+                <span id="submit-label">Get probabilities</span>
+                <svg class="submit-arrow" viewBox="0 0 20 20" fill="none" aria-hidden="true"><path d="M4 10h12m-5-5 5 5-5 5" stroke="currentColor" stroke-width="1.6" stroke-linecap="round" stroke-linejoin="round"/></svg>
+              </button>
+              <span class="keyboard-hint" id="keyboard-hint"><kbd>Ctrl</kbd> + <kbd>Enter</kbd></span>
+            </div>
+          </form>
+          <div class="editor-footnote">
+            <span>Examples are invented. Make them your own.</span>
+            <button type="button" class="text-button" id="reset">Reset example</button>
+          </div>
+        </section>
+
+        <section class="results-panel" id="results-panel" aria-labelledby="results-title" aria-busy="false">
+          <div class="panel-heading">
+            <h2 id="results-title">Probabilities</h2>
+            <span class="results-count" id="results-count" hidden></span>
+          </div>
+
+          <div class="empty-state" id="empty-state">
+            <div class="empty-mark" aria-hidden="true"><span></span><span></span><span></span></div>
+            <h3>Room for every possibility.</h3>
+            <p>Your results will appear here.<br>Run the example, or make it your own.</p>
+          </div>
+
+          <div class="loading-state" id="loading-state" hidden>
+            <div class="loading-orbit" aria-hidden="true"></div>
+            <h3 id="loading-title">Weighing the options</h3>
+            <p id="loading-detail">The model is reading your context.</p>
+            <span class="elapsed-timer" id="elapsed-timer" aria-hidden="true">0.0 s</span>
+          </div>
+
+          <div class="results-content" id="results-content" hidden>
+            <p class="result-question" id="result-question"></p>
+            <p class="stale-notice" id="stale-notice" role="status" hidden>Inputs changed. Run again to update these results.</p>
+            <ol class="probability-list" id="probability-list" aria-label="Answers ranked by probability"></ol>
+            <p class="probability-note">Probabilities are relative to the answers you provide.</p>
+            <div class="result-footer">
+              <div class="result-metadata">
+                <span class="result-model" id="result-model"></span>
+                <span id="result-timing"></span>
+              </div>
+              <button class="copy-button" id="copy-results" type="button">
+                <svg viewBox="0 0 18 18" fill="none" aria-hidden="true"><rect x="6" y="6" width="8" height="9" rx="1.5" stroke="currentColor" stroke-width="1.3"/><path d="M4 12H3.5A1.5 1.5 0 0 1 2 10.5v-7A1.5 1.5 0 0 1 3.5 2h6A1.5 1.5 0 0 1 11 3.5V4" stroke="currentColor" stroke-width="1.3"/></svg>
+                <span id="copy-label">Copy JSON</span>
+              </button>
+            </div>
+          </div>
+          <p class="sr-only" id="result-announcement" role="status" aria-live="polite" aria-atomic="true"></p>
+        </section>
+      </div>
+    </main>
+    <footer class="site-footer"><span>OpenSysOne</span><span>Context in. Possibilities out.</span></footer>
+  </div>
+  <noscript><p class="noscript-message">Enable JavaScript to load models and get probabilities.</p></noscript>
+</body>
+</html>

source/web/playground.js

@@ -0,0 +1,428 @@
+(() => {
+  "use strict";
+
+  const $ = (id) => document.getElementById(id);
+  const examples = [
+    {
+      text: "Ari and Noor are planning Saturday afternoon. Rain is forecast from noon onward. They want an indoor activity, enjoy art, and have a small budget. The city gallery has a free exhibition open until 6 pm. The hiking trail has no shelter, and the rooftop restaurant is expensive.",
+      question: "Which plan best fits their preferences?",
+      options: ["Visit the free art gallery", "Take a long hike", "Book a rooftop lunch"],
+    },
+    {
+      text: "Mina is expecting a parcel that needs a signature. The courier will arrive at 2 pm, but Mina will be away until 6 pm. Her building's staffed reception can sign for residents' parcels and stays open until 8 pm. Her neighbor Sam is away all day. Parcels may not be left outside.",
+      question: "Where should the courier leave Mina's parcel?",
+      options: ["With the staffed reception", "Outside Mina's door", "With her neighbor Sam"],
+    },
+    {
+      text: "Willow Studio offers a beginner pottery session on Saturday at 10 am for £25, an advanced ceramics session at 2 pm for £40, and a drawing class at 11 am for £20. Leo has never tried pottery and wants to learn it. He is free before 1 pm and can spend up to £30.",
+      question: "Which session should Leo book?",
+      options: ["The beginner pottery session", "The advanced ceramics session", "The drawing class"],
+    },
+  ];
+
+  let models = [];
+  let currentExample = 0;
+  let scoring = false;
+  let backendBusy = false;
+  let lastResponse = null;
+  let lastRequest = null;
+  let statusTimer = null;
+  let elapsedTimer = null;
+  let copyTimer = null;
+  let modelsLoading = false;
+  let statusSequence = 0;
+  let requestStarted = 0;
+  const fields = [$("context"), $("question"), $("options")];
+  const exampleButtons = Array.from(document.querySelectorAll("[data-example]"));
+
+  function optionsFromInput() {
+    return $("options").value.split(/\r?\n/).map((option) => option.trim()).filter(Boolean);
+  }
+
+  function requestFromInput() {
+    return {
+      model: $("model").value,
+      text: $("context").value.trim(),
+      question: $("question").value.trim(),
+      options: optionsFromInput(),
+    };
+  }
+
+  function updateCounts() {
+    const characters = $("context").value.length;
+    const count = optionsFromInput().length;
+    $("context-count").textContent = `${characters.toLocaleString()} character${characters === 1 ? "" : "s"}`;
+    $("option-count").textContent = `${count} option${count === 1 ? "" : "s"}`;
+  }
+
+  function setError(message, field) {
+    $("error-message").textContent = message || "";
+    $("error-message").hidden = !message;
+    fields.forEach((input) => input.removeAttribute("aria-invalid"));
+    if (field) {
+      field.setAttribute("aria-invalid", "true");
+      field.focus();
+    }
+  }
+
+  function connection(label, state) {
+    if ($("connection-text").textContent !== label) $("connection-text").textContent = label;
+    $("connection").dataset.state = state;
+  }
+
+  function syncControls() {
+    $("submit").disabled = scoring || backendBusy || !models.length || modelsLoading;
+    $("model").disabled = scoring || !models.length || modelsLoading;
+    $("submit").dataset.loading = String(scoring);
+    if (!scoring) $("submit-label").textContent = backendBusy ? "Model is busy" : "Get probabilities";
+    fields.forEach((input) => { input.readOnly = scoring; });
+    exampleButtons.forEach((button) => { button.disabled = scoring; });
+    $("clear").disabled = scoring;
+    $("reset").disabled = scoring;
+  }
+
+  function updateModel() {
+    const model = models.find((item) => item.id === $("model").value);
+    if (!model) return;
+    $("model-description").textContent = model.description || "Choose a context and compare possible answers.";
+    $("calibration-badge").hidden = false;
+    $("calibration-badge").textContent = model.calibrated ? "Calibrated checkpoint" : "Uncalibrated checkpoint";
+    $("calibration-badge").dataset.calibrated = String(Boolean(model.calibrated));
+    $("calibration-badge").title = model.calibrated
+      ? "This checkpoint includes a fitted probability calibration."
+      : "These probabilities have not been calibrated on a separate calibration set.";
+    $("model-limit").textContent = Number.isFinite(model.max_tokens) ? `${model.max_tokens.toLocaleString()}-token input limit` : "";
+    $("model-limit").title = "The limit includes the context, question and each possible answer.";
+  }
+
+  function updateStaleness() {
+    $("stale-notice").hidden = !lastRequest || JSON.stringify(requestFromInput()) === JSON.stringify(lastRequest);
+  }
+
+  function clearResults() {
+    lastResponse = null;
+    lastRequest = null;
+    $("probability-list").replaceChildren();
+    $("results-content").hidden = true;
+    $("results-count").hidden = true;
+    $("empty-state").hidden = false;
+    $("stale-notice").hidden = true;
+    $("copy-label").textContent = "Copy JSON";
+    $("result-announcement").textContent = "";
+    clearTimeout(copyTimer);
+  }
+
+  function loadExample(index) {
+    if (scoring) return;
+    const example = examples[index];
+    currentExample = index;
+    $("context").value = example.text;
+    $("question").value = example.question;
+    $("options").value = example.options.join("\n");
+    exampleButtons.forEach((button) => button.setAttribute("aria-pressed", String(Number(button.dataset.example) === index)));
+    setError("");
+    clearResults();
+    updateCounts();
+  }
+
+  async function jsonRequest(url, options = {}, timeout = 15000) {
+    const controller = new AbortController();
+    const timer = setTimeout(() => controller.abort(), timeout);
+    try {
+      const response = await fetch(url, { ...options, signal: controller.signal, cache: "no-store" });
+      let body;
+      try { body = await response.json(); } catch {
+        throw new Error(response.ok ? "The server returned an unreadable response. Please try again." : `The request failed (${response.status}). Please try again.`);
+      }
+      if (!response.ok) {
+        const error = new Error(typeof body.error === "string" ? body.error : `The request failed (${response.status}). Please try again.`);
+        error.status = response.status;
+        throw error;
+      }
+      return body;
+    } catch (error) {
+      if (error.name === "AbortError") throw new Error("The request timed out. The model may still be working; wait a moment and try again.");
+      if (error instanceof TypeError) throw new Error("Could not reach the model. Check the connection and try again.");
+      throw error;
+    } finally {
+      clearTimeout(timer);
+    }
+  }
+
+  async function loadModels() {
+    if (modelsLoading || scoring) return;
+    modelsLoading = true;
+    $("retry-models").hidden = true;
+    connection("Connecting", "connecting");
+    syncControls();
+    try {
+      const body = await jsonRequest("/api/models");
+      if (!Array.isArray(body.models)) throw new Error("The server did not return a model list. Please retry.");
+      models = body.models.filter((model) => model && typeof model.id === "string" && typeof model.label === "string");
+      const previous = $("model").value;
+      $("model").replaceChildren();
+      if (!models.length) {
+        const option = document.createElement("option");
+        option.textContent = "No models available";
+        option.value = "";
+        $("model").append(option);
+        $("model-description").textContent = "No checkpoints are currently available. Try again shortly.";
+        $("calibration-badge").hidden = true;
+        $("model-limit").textContent = "";
+        $("retry-models").hidden = false;
+        connection("No models available", "connecting");
+        return;
+      }
+      models.forEach((model) => {
+        const option = document.createElement("option");
+        option.value = model.id;
+        option.textContent = model.label;
+        $("model").append(option);
+      });
+      const requested = models.some((model) => model.id === previous) ? previous : body.default_model;
+      $("model").value = models.some((model) => model.id === requested) ? requested : models[0].id;
+      updateModel();
+      updateStaleness();
+      setError("");
+      connection("Ready", "ready");
+      void pollStatus();
+    } catch (error) {
+      models = [];
+      $("model").replaceChildren();
+      const option = document.createElement("option");
+      option.textContent = "Models unavailable";
+      option.value = "";
+      $("model").append(option);
+      $("model-description").textContent = "The model list could not be loaded.";
+      $("calibration-badge").hidden = true;
+      $("model-limit").textContent = "";
+      $("retry-models").hidden = false;
+      connection("Connection unavailable", "error");
+      setError(error.message);
+    } finally {
+      modelsLoading = false;
+      syncControls();
+    }
+  }
+
+  async function pollStatus() {
+    const sequence = ++statusSequence;
+    clearTimeout(statusTimer);
+    try {
+      const body = await jsonRequest("/api/status", {}, 7000);
+      if (sequence !== statusSequence) return;
+      backendBusy = body.status === "loading" || body.status === "scoring";
+      if (body.status === "loading") {
+        connection("Loading model", "busy");
+        if (scoring) {
+          $("loading-title").textContent = "Getting the model ready";
+          $("loading-detail").textContent = "The first request can take a little longer.";
+          $("submit-label").textContent = "Loading model…";
+        }
+      } else if (body.status === "scoring" || scoring) {
+        connection("Scoring", "busy");
+        if (scoring) {
+          $("loading-title").textContent = "Weighing the options";
+          $("loading-detail").textContent = "The model is comparing your possible answers.";
+          $("submit-label").textContent = "Scoring…";
+        }
+      } else if (body.status === "error") {
+        connection("Model needs a retry", "error");
+      } else {
+        connection(models.length ? "Ready" : "No models available", models.length ? "ready" : "connecting");
+      }
+    } catch {
+      if (sequence !== statusSequence) return;
+      if (!scoring) {
+        backendBusy = false;
+        connection("Connection unavailable", "error");
+      }
+    } finally {
+      if (sequence === statusSequence) {
+        syncControls();
+        statusTimer = setTimeout(pollStatus, scoring || backendBusy ? 1500 : 6000);
+      }
+    }
+  }
+
+  function percentage(value) {
+    const percent = value * 100;
+    if (percent > 0 && percent < .1) return "<0.1%";
+    return `${percent.toFixed(1)}%`;
+  }
+
+  function validateResponse(body, request) {
+    if (!body || !body.model || body.model.id !== request.model || typeof body.model.label !== "string"
+      || !Array.isArray(body.probabilities) || body.probabilities.length !== request.options.length) {
+      throw new Error("The model returned an incomplete response. Please try again.");
+    }
+    const remaining = new Set(request.options);
+    for (const row of body.probabilities) {
+      if (!row || !remaining.has(row.option) || typeof row.probability !== "number"
+        || !Number.isFinite(row.probability) || row.probability < 0 || row.probability > 1) {
+        throw new Error("The model returned invalid probabilities. Please try again.");
+      }
+      remaining.delete(row.option);
+    }
+    const total = body.probabilities.reduce((sum, row) => sum + row.probability, 0);
+    if (remaining.size || Math.abs(total - 1) > .001) throw new Error("The returned probabilities do not form a complete distribution. Please try again.");
+  }
+
+  function renderResults(body, request) {
+    lastResponse = body;
+    lastRequest = request;
+    clearTimeout(copyTimer);
+    $("copy-label").textContent = "Copy JSON";
+    $("probability-list").replaceChildren();
+    const ranked = body.probabilities.map((row) => ({ ...row, inputIndex: request.options.indexOf(row.option) }))
+      .sort((a, b) => b.probability - a.probability || a.inputIndex - b.inputIndex);
+    ranked.forEach((row) => {
+      const item = document.createElement("li");
+      item.className = "probability-item";
+      const label = document.createElement("div");
+      label.className = "probability-row";
+      const number = document.createElement("span");
+      number.className = "option-index";
+      number.textContent = String(row.inputIndex + 1).padStart(2, "0");
+      number.setAttribute("aria-label", `Input option ${row.inputIndex + 1}`);
+      const text = document.createElement("span");
+      text.className = "option-text";
+      text.textContent = row.option;
+      const probability = document.createElement("span");
+      probability.className = "option-probability";
+      probability.textContent = percentage(row.probability);
+      label.append(number, text, probability);
+      const track = document.createElement("div");
+      track.className = "probability-track";
+      track.setAttribute("aria-hidden", "true");
+      const fill = document.createElement("div");
+      fill.className = "probability-fill";
+      fill.style.width = `${row.probability * 100}%`;
+      track.append(fill);
+      item.append(label, track);
+      $("probability-list").append(item);
+    });
+    $("result-question").textContent = request.question;
+    $("results-count").textContent = `${ranked.length} options · ranked`;
+    $("results-count").hidden = false;
+    $("result-model").textContent = body.model.label;
+    const metadata = [];
+    if (Number.isFinite(body.elapsed_seconds)) metadata.push(`${body.elapsed_seconds.toFixed(2)} s`);
+    if (Number.isFinite(body.input_tokens)) metadata.push(`${body.input_tokens.toLocaleString()} input tokens`);
+    $("result-timing").textContent = metadata.join(" · ");
+    $("results-content").hidden = false;
+    $("empty-state").hidden = true;
+    updateStaleness();
+    $("result-announcement").textContent = `Results ready. Highest probability: ${ranked[0].option}, ${percentage(ranked[0].probability)}. ${ranked.length} options ranked.`;
+  }
+
+  async function score(event) {
+    if (event) event.preventDefault();
+    if (scoring || modelsLoading || !models.length) return;
+    if (backendBusy) { setError("The model is busy with another request. Please try again when it is ready."); return; }
+    const request = requestFromInput();
+    if (!request.text) { setError("Add some context for the model to read.", $("context")); return; }
+    if (!request.question) { setError("Add the question you want the model to answer.", $("question")); return; }
+    if (request.options.length < 2) { setError("Add at least two possible answers, one per line.", $("options")); return; }
+    if (new Set(request.options).size !== request.options.length) { setError("Each possible answer should be different. Remove any duplicate lines.", $("options")); return; }
+    setError("");
+    scoring = true;
+    requestStarted = performance.now();
+    $("results-panel").setAttribute("aria-busy", "true");
+    $("empty-state").hidden = true;
+    $("results-content").hidden = true;
+    $("results-count").hidden = true;
+    $("loading-state").hidden = false;
+    $("loading-title").textContent = "Weighing the options";
+    $("loading-detail").textContent = "The model is reading your context.";
+    $("submit-label").textContent = "Scoring…";
+    $("result-announcement").textContent = "Request started. Waiting for the model.";
+    const tick = () => { $("elapsed-timer").textContent = `${((performance.now() - requestStarted) / 1000).toFixed(1)} s`; };
+    tick();
+    elapsedTimer = setInterval(tick, 100);
+    syncControls();
+    connection("Scoring", "busy");
+    if (window.matchMedia("(max-width: 760px)").matches) {
+      $("results-panel").scrollIntoView({
+        behavior: window.matchMedia("(prefers-reduced-motion: reduce)").matches ? "auto" : "smooth",
+        block: "start",
+      });
+    }
+    void pollStatus();
+    try {
+      const body = await jsonRequest("/api/score", {
+        method: "POST", headers: { "Content-Type": "application/json" }, body: JSON.stringify(request),
+      }, 300000);
+      validateResponse(body, request);
+      renderResults(body, request);
+    } catch (error) {
+      setError(error.message);
+      if (error.status === 503) backendBusy = true;
+      $("results-content").hidden = !lastResponse;
+      $("results-count").hidden = !lastResponse;
+      $("empty-state").hidden = Boolean(lastResponse);
+      $("result-announcement").textContent = "The request could not be completed.";
+      updateStaleness();
+    } finally {
+      scoring = false;
+      clearInterval(elapsedTimer);
+      $("loading-state").hidden = true;
+      $("results-panel").setAttribute("aria-busy", "false");
+      syncControls();
+      void pollStatus();
+    }
+  }
+
+  async function copyResults() {
+    if (!lastResponse) return;
+    const text = JSON.stringify(lastResponse, null, 2);
+    let copied = false;
+    if (navigator.clipboard && window.isSecureContext) {
+      try { await navigator.clipboard.writeText(text); copied = true; } catch { /* Fall back for local HTTP. */ }
+    }
+    if (!copied) {
+      const field = document.createElement("textarea");
+      field.value = text;
+      field.className = "sr-only";
+      field.setAttribute("readonly", "");
+      document.body.append(field);
+      field.select();
+      try { copied = document.execCommand("copy"); } catch { copied = false; }
+      field.remove();
+      $("copy-results").focus();
+    }
+    $("copy-label").textContent = copied ? "Copied" : "Copy unavailable";
+    $("result-announcement").textContent = copied ? "Results JSON copied to the clipboard." : "Clipboard access is unavailable in this browser.";
+    clearTimeout(copyTimer);
+    copyTimer = setTimeout(() => { $("copy-label").textContent = "Copy JSON"; }, 2500);
+  }
+
+  $("decision-form").addEventListener("submit", score);
+  $("model").addEventListener("change", () => { updateModel(); updateStaleness(); setError(""); });
+  fields.forEach((input) => input.addEventListener("input", () => {
+    updateCounts(); updateStaleness(); setError("");
+    exampleButtons.forEach((button) => button.setAttribute("aria-pressed", "false"));
+  }));
+  exampleButtons.forEach((button) => button.addEventListener("click", () => loadExample(Number(button.dataset.example))));
+  $("reset").addEventListener("click", () => loadExample(currentExample));
+  $("clear").addEventListener("click", () => {
+    if (scoring) return;
+    fields.forEach((input) => { input.value = ""; });
+    exampleButtons.forEach((button) => button.setAttribute("aria-pressed", "false"));
+    setError(""); clearResults(); updateCounts(); $("context").focus();
+  });
+  $("retry-models").addEventListener("click", loadModels);
+  $("copy-results").addEventListener("click", copyResults);
+  document.addEventListener("keydown", (event) => {
+    if ((event.metaKey || event.ctrlKey) && event.key === "Enter") {
+      event.preventDefault();
+      if (!scoring) void score();
+    }
+  });
+  if (/Mac|iPhone|iPad/.test(navigator.platform || "")) $("keyboard-hint").textContent = "⌘ + Enter";
+  document.addEventListener("visibilitychange", () => {
+    if (!document.hidden && models.length) void pollStatus();
+  });
+  loadExample(0);
+  void loadModels();
+})();

sources/b05f7e3975a974c26ab667c6fdf91ec489a8cec0/manifest.json

@@ -0,0 +1,1155 @@
+{
+  "archive_sha256": "64a88e5e218f68e4b1dda00e7012eee0aaba4e4fcb9e0443b081c188528946c9",
+  "archive_size": 1654578,
+  "files": {
+    ".gitignore": {
+      "sha256": "c08a0eb968c43d53c1e54ef3f10e21ab4671c595cf25a2caa240b847acd06ddd",
+      "size": 73
+    },
+    "AGENTS.md": {
+      "sha256": "d27c6c6c5720db9c57efb1ee50a756c3d0486351fa2dcb33fd06a64fe7df7ac7",
+      "size": 1382
+    },
+    "FLEET_RUN.md": {
+      "sha256": "778339b8abd0dc1fa40d8c8b999a1cf39719a78cce8069bbf5b3f6bfa717b5be",
+      "size": 13851
+    },
+    "FLEET_SCOUT.md": {
+      "sha256": "789b020758a2713c74003f7d8d7af2d099c348d786c58c54136738a7ddcf7285",
+      "size": 8383
+    },
+    "HANDOVER.md": {
+      "sha256": "c46002891db5c557e28846a4f0adb89a531987c920da89a101493065b7bdfc24",
+      "size": 19142
+    },
+    "HF_MODEL_CARD.md": {
+      "sha256": "217a55e4c411423f93e12286c2b424d2ad886c3c5ae7f3092248bd08503dda5a",
+      "size": 6161
+    },
+    "HUGGINGFACE.md": {
+      "sha256": "521ae9303dfc69b65ba570811814c758434565e0b8057da97a135dcb5d7aab05",
+      "size": 5610
+    },
+    "JEV_HARNESS.md": {
+      "sha256": "c0f4a0d9055689b6a8be6fb66aadb43ed448609002618dd4d7e3857808c5e88b",
+      "size": 4683
+    },
+    "NEXT_STEPS.md": {
+      "sha256": "8fc19584153ca56b3e86ca77af7dd01e6d5a8d05a998973359dc003e5ff56a4f",
+      "size": 12181
+    },
+    "PLAN.md": {
+      "sha256": "9ec4049ed36158a8d176ac8f18b7ff16908ba5ab7f457188f7ccf2a4dfb5b820",
+      "size": 15930
+    },
+    "PLAYGROUND.md": {
+      "sha256": "506486e66e33e74925cd78092f02c0f50a1d163b2bcf5858bc52a2c4478352e5",
+      "size": 4790
+    },
+    "README.md": {
+      "sha256": "f5e5afde7c7073bd5f6e482e9f8eb61f8cd76268293a13626d6bc812a6bdc812",
+      "size": 3740
+    },
+    "RESEARCH_BRIEF.md": {
+      "sha256": "9ed40f7da30fbc0d1492d30e6bd5a2f315167f077293cbc3f7345de3a20ef164",
+      "size": 24894
+    },
+    "RESEARCH_NOTES.md": {
+      "sha256": "e11d9f1e09481e741cd333d7fe84e566bd468db40af77d92f5c361f4c428306c",
+      "size": 13713
+    },
+    "RESULTS.md": {
+      "sha256": "18d688c4e25f5e7008a9c6ae89da1e68370ec8b1d9c5b181f56908e4db22281d",
+      "size": 30956
+    },
+    "decision_model.py": {
+      "sha256": "a3d8aeb02a1ac765c6cc30ff175acad0664560f01ab5403e22cade924d17371e",
+      "size": 9258
+    },
+    "examples/jev_request.json": {
+      "sha256": "74f07501aa665284ab0611b6a3ed1fdc046821be10dddaa0a503efc52d7d7eb5",
+      "size": 703
+    },
+    "experiment.py": {
+      "sha256": "c036cbe9fbf6f71c1bb141dcfbecfb06a2f4bd697b070d2274b1931480e92031",
+      "size": 42981
+    },
+    "jev_harness.py": {
+      "sha256": "4d4e979cb7ae352bcdacaaa6d64045e6b5e550b1a9721d4bad545045bee6c67f",
+      "size": 16898
+    },
+    "playground.py": {
+      "sha256": "0eed2ab323dc73c7970b6d11fe41ddcb25e03f376af7e27891fd7e6c6a1e6e12",
+      "size": 18522
+    },
+    "results/20260916-fleet-setup/Qwen3-4B-Instruct-2507-cdbee75f-files.json": {
+      "sha256": "3ceddf4e5228246ddb9a6ad381fb31793a61b050a8fb10934a345225e6068fe6",
+      "size": 1971
+    },
+    "results/20260916-fleet-setup/Qwen3.5-2B-15852e8c-files.json": {
+      "sha256": "a78fb2654833d89f5b274b562a4595f2dff001065db2cf64f83bcd1a206c7576",
+      "size": 1950
+    },
+    "results/20260916-fleet-setup/cpu-tests.json": {
+      "sha256": "2110698e40353ee830b3f7bf792c5f2c8ea0aff54c8fa7d9b6b49fde99540b71",
+      "size": 328
+    },
+    "results/20260916-fleet-setup/crossfit-tests.json": {
+      "sha256": "a804cfb5eaf7a963085cb399da2f0789f8a50c56bd6c38f44b7906c4fffe2d3f",
+      "size": 532
+    },
+    "results/20260916-fleet-setup/evidence-index.json": {
+      "sha256": "db4d25099b45af53cffa54551cb0c56f82b1a0228915455873183f31a228148e",
+      "size": 12480
+    },
+    "results/20260916-fleet-setup/fleet-current-status.json": {
+      "sha256": "06b07b3ed301bdfaacfe4ded988033e525e4ea7580b0f6bdbe6bdaceb4a1bab9",
+      "size": 11377
+    },
+    "results/20260916-fleet-setup/fleet-launch-verification.json": {
+      "sha256": "9f6ee33c3004645c851033f87b8fb748478d37a2b59dff75b35fff9e1f45a413",
+      "size": 553
+    },
+    "results/20260916-fleet-setup/fleet-plan.json": {
+      "sha256": "bff9ed43f7fd7710ec3915144424daf1f0ffb1160ae86ee3a654cbba7974cd4b",
+      "size": 1795
+    },
+    "results/20260916-fleet-setup/fleet-state.json": {
+      "sha256": "a84107d6c7ee5c96995cd3fd314d5ce217c1ef29f6296f8771b3bdfdaa6fc018",
+      "size": 664
+    },
+    "results/20260916-fleet-setup/gx10-current-best_validation_predictions.json": {
+      "sha256": "9f1f8ccbc26a29ba0a0615ff1cbf278776c01c003b63176bae41df5a5bde92e6",
+      "size": 332042
+    },
+    "results/20260916-fleet-setup/gx10-current-best_validation_selection.json": {
+      "sha256": "9cbc9717355d4d8e17d1d072cae9c243c37e2497e691ca0685fe5b6194c3d74b",
+      "size": 19696
+    },
+    "results/20260916-fleet-setup/gx10-current-correctness_initial.json": {
+      "sha256": "5a3a0b5b7cd5b9ac2840fe9a6de37d60fc069f1958e315a8a0c973799a960198",
+      "size": 352
+    },
+    "results/20260916-fleet-setup/gx10-current-manifest.json": {
+      "sha256": "8c81e2582258ec2c554b284e355e8059f317d91d24ec775e32dc86dc696c5845",
+      "size": 13020
+    },
+    "results/20260916-fleet-setup/gx10-current-startup-verification.json": {
+      "sha256": "875e7c410ad0c98328737c2f34b5da1d9dedf9d5ba44cfed0b9bdd43226f73af",
+      "size": 24964
+    },
+    "results/20260916-fleet-setup/gx10-current-validation_step_000178_predictions.json": {
+      "sha256": "9f1f8ccbc26a29ba0a0615ff1cbf278776c01c003b63176bae41df5a5bde92e6",
+      "size": 332042
+    },
+    "results/20260916-fleet-setup/gx10-python-stack-proof.json": {
+      "sha256": "9ee3222204613414ecbcba0a2faecf740c03e1d771fbdcc0af5ff5d98a5bb62a",
+      "size": 364
+    },
+    "results/20260916-fleet-setup/gx10-resume-state-verification.json": {
+      "sha256": "5ce557a67fd757854667e36473754c6b14928615c9a6564bc036fe3bd0e1465c",
+      "size": 635
+    },
+    "results/20260916-fleet-setup/gx10-transition-checkpoint.json": {
+      "sha256": "f14a3b2da291655c3296e1a5fad439c83de1cf7e143318e74781b4fb1767f6a6",
+      "size": 581
+    },
+    "results/20260916-fleet-setup/selection-diagnostic.json": {
+      "sha256": "783190fe970f0930b449632065b2689d5b2e91a6d7a31415cf4710c9907ba55f",
+      "size": 4482
+    },
+    "results/20260916-fleet-setup/selection_migration.json": {
+      "sha256": "37bc6f9606d3e204bcdf45e0cd50bea2e5713f8d639c98fbd39f92c665c527bc",
+      "size": 1389
+    },
+    "results/20260916-fleet-setup/spark-a-best-validation-selection.json": {
+      "sha256": "e6cd9400edd4d510fe0342f726aead056c075c0829a767400e1ffd335dc86b39",
+      "size": 19697
+    },
+    "results/20260916-fleet-setup/spark-a-campaign-plan.json": {
+      "sha256": "f04bbd72d5a7cc59980714f1ef674ec7c9af309d8a956a7e5a03b766c647af74",
+      "size": 777
+    },
+    "results/20260916-fleet-setup/spark-a-campaign-state-snapshot.json": {
+      "sha256": "417d9f1552cfa82ddc8ad49224b986ef5d58f89d942a73b338a6aa6646cb4ab9",
+      "size": 1421
+    },
+    "results/20260916-fleet-setup/spark-a-campaign-updates-verified.json": {
+      "sha256": "52eae7f877102c4c46a9cf3275a354784f0c95f13c9772cc578db12c101a7698",
+      "size": 4674
+    },
+    "results/20260916-fleet-setup/spark-a-candidate-launch.json": {
+      "sha256": "e35162501a6d207f7467b44cc27b3d3fb1133db4f212a0b84942271909bdda91",
+      "size": 815
+    },
+    "results/20260916-fleet-setup/spark-a-inherited-validation-selection.json": {
+      "sha256": "2659b2b3fe8b7c8d9e614b30c25a99ad7100cd007892fba9b4ea6629fde0d887",
+      "size": 19687
+    },
+    "results/20260916-fleet-setup/spark-a-initial-validation-selection.json": {
+      "sha256": "e6cd9400edd4d510fe0342f726aead056c075c0829a767400e1ffd335dc86b39",
+      "size": 19697
+    },
+    "results/20260916-fleet-setup/spark-a-inputs-verified.json": {
+      "sha256": "7e6e8642a752993aa7181dcfb96e016ca4dbdf454109ff99d7174efd5a3c6f5b",
+      "size": 2202
+    },
+    "results/20260916-fleet-setup/spark-a-launcher.json": {
+      "sha256": "422197489b6de521d73d746a8bf21cfabc9ac82da82dfa892cfd68ce3526fd34",
+      "size": 140
+    },
+    "results/20260916-fleet-setup/spark-a-model-copy-verified.json": {
+      "sha256": "ffed13b2c1945c091bf2ea586556f94ba5e1ff67e7aeb92e1cbd49c488ae2ef9",
+      "size": 2188
+    },
+    "results/20260916-fleet-setup/spark-a-pilot-updates-verified.json": {
+      "sha256": "0ba2ad5967dbc5fc2e7a01a4bbf7a4ec6f1a7fe50bb815d9cefc63f314d1b154",
+      "size": 2897
+    },
+    "results/20260916-fleet-setup/spark-a-pilot/best_validation_predictions.json": {
+      "sha256": "e671e1508185765552b0f933ba03f356be62143c531d8ef534457d34b1645c9b",
+      "size": 333136
+    },
+    "results/20260916-fleet-setup/spark-a-pilot/correctness_final.json": {
+      "sha256": "cb2649db1735dbcbc90f83d9f95520a7dd553471b7d0c88316205207ad66acb8",
+      "size": 355
+    },
+    "results/20260916-fleet-setup/spark-a-pilot/correctness_initial.json": {
+      "sha256": "5d6334a0c66ed567fa9709df4ea5b6040c88bb5258e864332e88be2bb9586634",
+      "size": 355
+    },
+    "results/20260916-fleet-setup/spark-a-pilot/data_filter.json": {
+      "sha256": "3a514e34a8a6f7e35776bafe5b19015d6c5cec0eaabea5558aa896471aba3b56",
+      "size": 2360
+    },
+    "results/20260916-fleet-setup/spark-a-pilot/initial_validation_predictions.json": {
+      "sha256": "e671e1508185765552b0f933ba03f356be62143c531d8ef534457d34b1645c9b",
+      "size": 333136
+    },
+    "results/20260916-fleet-setup/spark-a-pilot/manifest.json": {
+      "sha256": "18672b6c49227a08ba1303a4958a9fa08ae78b5fe59e30c99902e5d61ccaddf0",
+      "size": 12238
+    },
+    "results/20260916-fleet-setup/spark-a-pilot/summary.json": {
+      "sha256": "9bda8ed53ed42f7a8b4dea08acbb6c4c58635bf64dafa925b1a3575e9a4f89b4",
+      "size": 11499
+    },
+    "results/20260916-fleet-setup/spark-a-pilot/training.jsonl": {
+      "sha256": "fe3be4437a159b98ddc84fb2da04f42de44aa10eebd8e19029d92f4a9585bc8f",
+      "size": 2082
+    },
+    "results/20260916-fleet-setup/spark-a-pilot/validation.jsonl": {
+      "sha256": "ef002be6c6dd5a27fa35f555d0a3fb3544f885ed806e026edeb2491fbb9ccd84",
+      "size": 6191
+    },
+    "results/20260916-fleet-setup/spark-a-pilot/validation_step_000008_predictions.json": {
+      "sha256": "d520b46127014161384c46d856fcfa167d4bb667a369235678f2e96ab0e54029",
+      "size": 333027
+    },
+    "results/20260916-fleet-setup/spark-a-python-stack-verified.json": {
+      "sha256": "8391cccb0d72f8a9af7e631075435a83ba3582c8bdd30b6ead0a3e50d90c46d3",
+      "size": 1965
+    },
+    "results/20260916-fleet-setup/spark-a-resume-state-verified.json": {
+      "sha256": "18ad5ae7e27f1d7c20b4de4cc0537fd1d77cf665f00ec91c87df9872c328d3c6",
+      "size": 1008
+    },
+    "results/20260916-fleet-setup/spark-a-resume-verified.json": {
+      "sha256": "7bbd348d2aed3478ec8967913068ce50c85acc6853d2dd33c56e1772c26369f0",
+      "size": 625
+    },
+    "results/20260916-fleet-setup/spark-a-service-stop.json": {
+      "sha256": "b2091baa22a19691fd9e5f4ce17c4bf61bcc07f2bdecd67b42fcd8e22b9af907",
+      "size": 1923
+    },
+    "results/20260916-fleet-setup/spark-a-setup-status.json": {
+      "sha256": "9ebe7f313f592080899534e55fda097aece13e559bf6fde67a0489f191cbe50d",
+      "size": 102
+    },
+    "results/20260916-fleet-setup/spark-a-training-correctness-initial.json": {
+      "sha256": "cb2649db1735dbcbc90f83d9f95520a7dd553471b7d0c88316205207ad66acb8",
+      "size": 355
+    },
+    "results/20260916-fleet-setup/spark-a-training-manifest.json": {
+      "sha256": "839a92247fa8a861e853e46a18a97cebb9b4514b327562dc59066b8363fdf06a",
+      "size": 12984
+    },
+    "results/20260916-fleet-setup/spark-a-verification/data_filter.json": {
+      "sha256": "3a514e34a8a6f7e35776bafe5b19015d6c5cec0eaabea5558aa896471aba3b56",
+      "size": 2360
+    },
+    "results/20260916-fleet-setup/spark-a-verification/http_255_choices_response.json": {
+      "sha256": "36ce43c4fc4885a73daa7f57c27143f311d0a075621e8d2f1121a4cf64b815f4",
+      "size": 12050
+    },
+    "results/20260916-fleet-setup/spark-a-verification/http_long_context_response.json": {
+      "sha256": "762a8383eaec4cc0ca22738d8cc9abde026781c37468ec2b5bf3058fd0e487d5",
+      "size": 216
+    },
+    "results/20260916-fleet-setup/spark-a-verification/http_response.json": {
+      "sha256": "f43f195c9ac3b62714e4dc11f99b71ac7e56aec89f979a481298b8039bd7a24f",
+      "size": 859
+    },
+    "results/20260916-fleet-setup/spark-a-verification/manifest.json": {
+      "sha256": "6a598e188f2d829ac3b3f7762fd67047691ba2246bddf959e3a073b8640986f7",
+      "size": 264
+    },
+    "results/20260916-fleet-setup/spark-a-verification/reload_predictions.json": {
+      "sha256": "f56623fbe5d0c26e6433bea78c77ecd61f4459f1ae3b0b62b808dae4a836b2fd",
+      "size": 10480
+    },
+    "results/20260916-fleet-setup/spark-a-verification/stress.json": {
+      "sha256": "5471c772bf1f56ad0c2df20c329eccf77aa0161eeb5953295c4959c9e298cf06",
+      "size": 1169
+    },
+    "results/20260916-fleet-setup/spark-a-verification/verification.json": {
+      "sha256": "3835333780af23e0c3398aa22b4b0da207a399e26ffa868978a457c4370a7f02",
+      "size": 2300
+    },
+    "results/20260916-fleet-setup/spark-a-warmstart-verified.json": {
+      "sha256": "516d935684a1a747ea6d74389568507a2fe49743a4d5b11dca89524f5992eaaf",
+      "size": 634
+    },
+    "results/20260916-fleet-setup/spark-b-best-validation-selection.json": {
+      "sha256": "9d16f6144a3ff81ffbd1a6cf78078d4bd01fa683f8bbfd4028627591686958f6",
+      "size": 19690
+    },
+    "results/20260916-fleet-setup/spark-b-correctness-initial.json": {
+      "sha256": "39d23512e99743437b88b5b10110e501853976b5292581481b5d2a5b7f2c5cdd",
+      "size": 391
+    },
+    "results/20260916-fleet-setup/spark-b-current-status.json": {
+      "sha256": "00af691a392d04dc2b8359a4232f836ec8f96501153025acc91ff370bcc9f357",
+      "size": 2861
+    },
+    "results/20260916-fleet-setup/spark-b-data-filter.json": {
+      "sha256": "e60da06955fe9e8a70a3de1ddaeb261e53af8f3a5e1d13ee62932d3773a1be24",
+      "size": 1572
+    },
+    "results/20260916-fleet-setup/spark-b-evidence-index.json": {
+      "sha256": "8b6e4a1172667c614749a3135d060183eefbf701a4f0c011a57b452d0d9b4490",
+      "size": 6279
+    },
+    "results/20260916-fleet-setup/spark-b-http-255-choices-response.json": {
+      "sha256": "fe9c590782a9584a7ccca84efcce728918e6a0d53edb11a86de4ca06a9d0ec9c",
+      "size": 11977
+    },
+    "results/20260916-fleet-setup/spark-b-http-long-context-response.json": {
+      "sha256": "77537e1b07b9746e7c75c503de97d570bbedfdc10c4e3be602d9aca7dfa6fd98",
+      "size": 202
+    },
+    "results/20260916-fleet-setup/spark-b-http-response.json": {
+      "sha256": "2baaeafd62463d40bd7318f0a24bfa423897f1139de2c60ccd16d072fd4955c6",
+      "size": 844
+    },
+    "results/20260916-fleet-setup/spark-b-inherited-validation-selection.json": {
+      "sha256": "9d16f6144a3ff81ffbd1a6cf78078d4bd01fa683f8bbfd4028627591686958f6",
+      "size": 19690
+    },
+    "results/20260916-fleet-setup/spark-b-initial-validation-selection.json": {
+      "sha256": "9d16f6144a3ff81ffbd1a6cf78078d4bd01fa683f8bbfd4028627591686958f6",
+      "size": 19690
+    },
+    "results/20260916-fleet-setup/spark-b-model-copy-verified.json": {
+      "sha256": "d6c5d97275d80f93e0a0713f3801eed6b173046d9d5c95989fa91504ad5c5b3f",
+      "size": 2155
+    },
+    "results/20260916-fleet-setup/spark-b-plan.json": {
+      "sha256": "509505b3fcb2af1d50b16bfdecd5833a7691da217700923d080aced6d679016f",
+      "size": 777
+    },
+    "results/20260916-fleet-setup/spark-b-python-stack-verified.json": {
+      "sha256": "c1eecf2c6d3c1a664180abdc99161a2e0372af5b0dbfab4fc354ebb49a5abfb6",
+      "size": 1965
+    },
+    "results/20260916-fleet-setup/spark-b-reload-predictions.json": {
+      "sha256": "4cc860ca3e6a65a904d17609d74102e39283480fd4f1e5f69cee4d8c7dc135bb",
+      "size": 10455
+    },
+    "results/20260916-fleet-setup/spark-b-service-stop.json": {
+      "sha256": "8706294be28526fc6fb1ece61964a61085a32636e5cd399387a2bd6e3dd7012f",
+      "size": 1201
+    },
+    "results/20260916-fleet-setup/spark-b-setup-launch.json": {
+      "sha256": "a065853c9ad9d6f63475fcc4996b289c7bb914bf50129e06b38d820145499347",
+      "size": 915
+    },
+    "results/20260916-fleet-setup/spark-b-setup-stages.log": {
+      "sha256": "13216b31b4287dc33421e7a8264133b5615a15e463b45fffa740eacfd4dd843d",
+      "size": 260
+    },
+    "results/20260916-fleet-setup/spark-b-setup-status.json": {
+      "sha256": "3f11175e52779f7ffb1e1848f6a57e859d1a318b10c8424d2695464d060a1a82",
+      "size": 102
+    },
+    "results/20260916-fleet-setup/spark-b-startup-verification.json": {
+      "sha256": "9eeef85649a5765a9a465fd082d5ef71687afbe1c7c910b6b1395507b3faeb94",
+      "size": 9581
+    },
+    "results/20260916-fleet-setup/spark-b-stress.json": {
+      "sha256": "617709fb02a0098b6ccbfce17d5817cbacf680178cd0a2d50e6c02557b2160d8",
+      "size": 1162
+    },
+    "results/20260916-fleet-setup/spark-b-training-manifest.json": {
+      "sha256": "e6f39dc5c797a0c1f5aa6ca176509d4cb1ef176a98af0bb6ee0946a5acbbc09e",
+      "size": 10911
+    },
+    "results/20260916-fleet-setup/spark-b-verification-manifest.json": {
+      "sha256": "891e5a059daa63397bcd403bf4f2e4297c5aba983272d25b316ac6a68c3b9457",
+      "size": 264
+    },
+    "results/20260916-fleet-setup/spark-b-verification.json": {
+      "sha256": "0ee15bcf7d01d312b91ec935fefa9944338847cf8a1f67cabba79e1a141c0af3",
+      "size": 2279
+    },
+    "results/20260916T154714Z/base_token_yes_minus_no_predictions.json": {
+      "sha256": "5c6dc5ab1c21d68cff293fbf73cdea66447d9a7ad2c3aaaac976608c7a696ecd",
+      "size": 31072
+    },
+    "results/20260916T154714Z/calibrated_predictions.json": {
+      "sha256": "cfd85bd7ae72ac090f326a8fe534b242230dad9b8772ddb50d3bddda50e25e70",
+      "size": 33872
+    },
+    "results/20260916T154714Z/calibration.jsonl": {
+      "sha256": "09e51eeffa7cdc2faad78f3dc6fd7b994580aae7e0ac2ecf8f1ceafdd7f984a0",
+      "size": 16193
+    },
+    "results/20260916T154714Z/calibration_predictions.json": {
+      "sha256": "f0d1fb5ba5d9778ccc4db37830372ec2257ce76f6a082eb63b00ee6009f768be",
+      "size": 23027
+    },
+    "results/20260916T154714Z/exit_code": {
+      "sha256": "4355a46b19d348dc2f57c046f8ef63d4538ebb936000f3c9ee954a27460dd865",
+      "size": 2
+    },
+    "results/20260916T154714Z/initial_scalar_predictions.json": {
+      "sha256": "6aa8afae1afb92d4ebf273d80741f3b165e694a20280a1da2e9144425c2d8294",
+      "size": 28395
+    },
+    "results/20260916T154714Z/manifest.json": {
+      "sha256": "4b4a55380d16561b0703d2b6c1d94b05a7f4d129b193656c1003dc80548e6414",
+      "size": 3462
+    },
+    "results/20260916T154714Z/parity_diagnosis.json": {
+      "sha256": "060e44827c102ed8c4f5a317dea9822a94fa8d87d8d795fb4a596e9954947caa",
+      "size": 6482
+    },
+    "results/20260916T154714Z/run.log": {
+      "sha256": "1659885bb673aed6f5ad0249111c75edab218bb46aefa7e2d3c38f2bc381448c",
+      "size": 11626
+    },
+    "results/20260916T154714Z/test.jsonl": {
+      "sha256": "4c0c365566099ff0941b2da473d6d3cca445ef1174c931d9d93f9bb475679599",
+      "size": 22779
+    },
+    "results/20260916T154714Z/train.jsonl": {
+      "sha256": "5ff456f912cf22ddefd8e0e24baf3956f0f010cd4c605e186a36659b4b931ea0",
+      "size": 61325
+    },
+    "results/20260916T154714Z/trained_predictions.json": {
+      "sha256": "8ebd0aad15bd04b4399cb467c4bad2bca813345179f737fbe510031d05a2e8a8",
+      "size": 34043
+    },
+    "results/20260916T154714Z/training.json": {
+      "sha256": "e3b0a701faa2eb67a3bbb70f4662e3f71d1587a396443afa709ec5e548577723",
+      "size": 10673
+    },
+    "results/20260916T155124Z/base_token_yes_minus_no_predictions.json": {
+      "sha256": "14bdc2ffad677e3fd3b6c58fae74efe3a7530df29ac9e181c2379a0cd5085ce1",
+      "size": 33837
+    },
+    "results/20260916T155124Z/benchmark.json": {
+      "sha256": "01027c6338ff2e36e3b7808f47c4f279bc9a2e52f24efc0bf7fd5744b5b572d8",
+      "size": 7692
+    },
+    "results/20260916T155124Z/calibrated_predictions.json": {
+      "sha256": "807e1f999a9ee64fa5f93851d1ee29da2d0fd41ed23e44e46f6391fb1f822a2c",
+      "size": 33865
+    },
+    "results/20260916T155124Z/calibration.jsonl": {
+      "sha256": "09e51eeffa7cdc2faad78f3dc6fd7b994580aae7e0ac2ecf8f1ceafdd7f984a0",
+      "size": 16193
+    },
+    "results/20260916T155124Z/calibration_predictions.json": {
+      "sha256": "1f3a0277c66f27c5ff08028ce6a7ccd90b9ccd5f7105b74fb8a3ae69c532422d",
+      "size": 23026
+    },
+    "results/20260916T155124Z/correctness.json": {
+      "sha256": "566f735523c8d1e46e4b11534b436d11edf1a9b1aef90e380aa10024b2b19abe",
+      "size": 329
+    },
+    "results/20260916T155124Z/exit_code": {
+      "sha256": "9a271f2a916b0b6ee6cecb2426f0b3206ef074578be55d9bc94f6f3fe3ab86aa",
+      "size": 2
+    },
+    "results/20260916T155124Z/initial_scalar_predictions.json": {
+      "sha256": "6aa8afae1afb92d4ebf273d80741f3b165e694a20280a1da2e9144425c2d8294",
+      "size": 28395
+    },
+    "results/20260916T155124Z/manifest.json": {
+      "sha256": "51b417c25f4b6f169b214bd6e4662a59b59365fc7e048feaa555941a05208a61",
+      "size": 3609
+    },
+    "results/20260916T155124Z/metrics.json": {
+      "sha256": "f5ecee19932286363c1e95bdb91c977dc907c9dc2a0f2bee85cfbb9c01472738",
+      "size": 6152
+    },
+    "results/20260916T155124Z/run.log": {
+      "sha256": "e0589922a98bdf412f749004cae387f8f60ef9c6f43f983de2cf433f6d41b6bc",
+      "size": 17478
+    },
+    "results/20260916T155124Z/test.jsonl": {
+      "sha256": "4c0c365566099ff0941b2da473d6d3cca445ef1174c931d9d93f9bb475679599",
+      "size": 22779
+    },
+    "results/20260916T155124Z/train.jsonl": {
+      "sha256": "5ff456f912cf22ddefd8e0e24baf3956f0f010cd4c605e186a36659b4b931ea0",
+      "size": 61325
+    },
+    "results/20260916T155124Z/trained_predictions.json": {
+      "sha256": "a0a020c3ec7134e82eab3fe22fe2e533d0dfe2393bf2e2f21c26020c09e5ba2a",
+      "size": 34016
+    },
+    "results/20260916T155124Z/training.json": {
+      "sha256": "6586a74826b7c0c7423a673cb7d9f6ce753dd097513802da18f975d769c9620a",
+      "size": 10643
+    },
+    "results/20260916T155314Z/benchmark.json": {
+      "sha256": "74fe266abe96c407487ccbe71ad4d30b78990b8c3382b68f64063ce76f85ee28",
+      "size": 7688
+    },
+    "results/20260916T155314Z/calibrated_predictions.json": {
+      "sha256": "aa4ffbf0fd363d2ba5a7f6d4ccff355dc636e88f8ef1b723eb2b0426e73aac46",
+      "size": 33876
+    },
+    "results/20260916T155314Z/calibration.jsonl": {
+      "sha256": "09e51eeffa7cdc2faad78f3dc6fd7b994580aae7e0ac2ecf8f1ceafdd7f984a0",
+      "size": 16193
+    },
+    "results/20260916T155314Z/calibration_predictions.json": {
+      "sha256": "eb2b47919f765cef2eb6d38ea1aff23e17ec2691f3b8052ee205b126130d7905",
+      "size": 22993
+    },
+    "results/20260916T155314Z/correctness.json": {
+      "sha256": "faf4077e01a6b0fede1988c98a586d71fc81c454b621a8159f9206d097aacce3",
+      "size": 335
+    },
+    "results/20260916T155314Z/exit_code": {
+      "sha256": "9a271f2a916b0b6ee6cecb2426f0b3206ef074578be55d9bc94f6f3fe3ab86aa",
+      "size": 2
+    },
+    "results/20260916T155314Z/manifest.json": {
+      "sha256": "51d594b45c7d67db6ea0be5a61d9fdf1d3d8b7e2b6805630e0a5bb255b734761",
+      "size": 3725
+    },
+    "results/20260916T155314Z/metrics.json": {
+      "sha256": "f0d51ea2136aff729163daccd56435636b751633a826f1e834cd4b332fb8e880",
+      "size": 4966
+    },
+    "results/20260916T155314Z/resume_verification.json": {
+      "sha256": "e9e80d9357d17c01f650f9b42cd358ed7193588736c98d91aaed21bc9ed30122",
+      "size": 773
+    },
+    "results/20260916T155314Z/resumed_initial_predictions.json": {
+      "sha256": "a0a020c3ec7134e82eab3fe22fe2e533d0dfe2393bf2e2f21c26020c09e5ba2a",
+      "size": 34016
+    },
+    "results/20260916T155314Z/run.log": {
+      "sha256": "05e8ef6eb53ff8fb01fb2571054501a4fd70b8d301e1dd35c81b248c02eb5769",
+      "size": 6914
+    },
+    "results/20260916T155314Z/test.jsonl": {
+      "sha256": "4c0c365566099ff0941b2da473d6d3cca445ef1174c931d9d93f9bb475679599",
+      "size": 22779
+    },
+    "results/20260916T155314Z/train.jsonl": {
+      "sha256": "5ff456f912cf22ddefd8e0e24baf3956f0f010cd4c605e186a36659b4b931ea0",
+      "size": 61325
+    },
+    "results/20260916T155314Z/trained_predictions.json": {
+      "sha256": "bdd004b8bb553cd446f159491f2be9c9e5e3f8f9b92061454ff20fb9850fcc92",
+      "size": 34004
+    },
+    "results/20260916T155314Z/training.json": {
+      "sha256": "6d56c059831aff521d59894702d7a66574d9e7ba8f965f745fa29e7531c5eb23",
+      "size": 180
+    },
+    "results/20260916T161253Z-precision/exit_code": {
+      "sha256": "9a271f2a916b0b6ee6cecb2426f0b3206ef074578be55d9bc94f6f3fe3ab86aa",
+      "size": 2
+    },
+    "results/20260916T161253Z-precision/manifest.json": {
+      "sha256": "7626809df6622a05507b11e65444e1d1b28e9a03ea81838169977991aabc6de5",
+      "size": 1169
+    },
+    "results/20260916T161253Z-precision/precision.json": {
+      "sha256": "5151de2f07119b2c6de007060899883f0cdb36ff728982df2b95de092d0097d8",
+      "size": 110342
+    },
+    "results/20260916T161253Z-precision/run.log": {
+      "sha256": "215e898ab1cf849c98ef56263e8d22a570d79263af2a39af925863adc394591f",
+      "size": 1655
+    },
+    "results/20260916T161355Z-precision/exit_code": {
+      "sha256": "9a271f2a916b0b6ee6cecb2426f0b3206ef074578be55d9bc94f6f3fe3ab86aa",
+      "size": 2
+    },
+    "results/20260916T161355Z-precision/manifest.json": {
+      "sha256": "93d8421244498c398f437adf4e6af0b4eb6b9b1a6f4d447564d1442faba56eae",
+      "size": 1103
+    },
+    "results/20260916T161355Z-precision/precision.json": {
+      "sha256": "d53276f38ba72bb01219d5e70183476cded0cd1b4bddb4cfb1ddddb9ec212000",
+      "size": 153559
+    },
+    "results/20260916T161355Z-precision/run.log": {
+      "sha256": "6696cda7c457d6c5c178d3adc4301b254dfe406b22dcfa8f0975c8b055755097",
+      "size": 2359
+    },
+    "results/20260916T182352Z-train/correctness_final.json": {
+      "sha256": "55e085c42e416d00fd4cdad9fb72d121f780fb9005226549ce7eaeddc9eea051",
+      "size": 408
+    },
+    "results/20260916T182352Z-train/correctness_initial.json": {
+      "sha256": "7414b27fe26d2d52ea023ba5cc4d8bc9a46d9bf60a8d54358ad2ad803c03c26a",
+      "size": 479
+    },
+    "results/20260916T182352Z-train/data_filter.json": {
+      "sha256": "e60da06955fe9e8a70a3de1ddaeb261e53af8f3a5e1d13ee62932d3773a1be24",
+      "size": 1572
+    },
+    "results/20260916T182352Z-train/initial_validation_predictions.json": {
+      "sha256": "4ee92802d7a8d642c8801dfd96f5bb18df80f745632fc1c8d99f9da9fe731c5b",
+      "size": 332151
+    },
+    "results/20260916T182352Z-train/manifest.json": {
+      "sha256": "e1efa2f4d350c678ba726e88dc85d8e7f3c8dc4e698742897892e578e5d94c58",
+      "size": 9460
+    },
+    "results/20260916T182352Z-train/resume_verification.json": {
+      "sha256": "501cb2750d0d5ba914dd066bd1312b39ec1e476ab8c074ac1b2758680f8a1928",
+      "size": 986
+    },
+    "results/20260916T182352Z-train/summary.json": {
+      "sha256": "05cf399210326916c92e404ed815ca8644bd3db09fa027cbfe4686fac47a902f",
+      "size": 11569
+    },
+    "results/20260916T182352Z-train/training.jsonl": {
+      "sha256": "231253f8237df8a40c95e6550fc36975b12b3b657e80da72c98a7712d365e48d",
+      "size": 10396
+    },
+    "results/20260916T182352Z-train/validation.jsonl": {
+      "sha256": "fdc0ad4f5d854f2ea33e45e644117e16d831d01576b9f011ccbca75745bc6c9b",
+      "size": 6202
+    },
+    "results/20260916T182352Z-train/validation_step_000040_predictions.json": {
+      "sha256": "d44337f6e32bd8f1e9db01f533ce031dafec5a6162fe43d19dde40e0a3c39c75",
+      "size": 332208
+    },
+    "results/20260916T183240Z-train/correctness_final.json": {
+      "sha256": "f941e3b1c39b5636b59ac513a9a0673ef683d708c52e505cb95b8271163efaf8",
+      "size": 391
+    },
+    "results/20260916T183240Z-train/correctness_initial.json": {
+      "sha256": "39d23512e99743437b88b5b10110e501853976b5292581481b5d2a5b7f2c5cdd",
+      "size": 391
+    },
+    "results/20260916T183240Z-train/manifest.json": {
+      "sha256": "ea6f9690fed03a43d2698161c3aae1685aca289cffd7f430e4ade5c32d16bb2d",
+      "size": 9818
+    },
+    "results/20260916T183240Z-train/summary.json": {
+      "sha256": "873e71eda53e3d000d805faa86d5a436611e97d74f331f89b9ed08d055f116a4",
+      "size": 11522
+    },
+    "results/20260916T183240Z-train/training.jsonl": {
+      "sha256": "7ea9596dd9503513d55715f4d0dee5f3aa8e22bd291288c4fde32511f8691677",
+      "size": 261
+    },
+    "results/20260916T183240Z-train/validation.jsonl": {
+      "sha256": "5f3e0dc2ed0e96ebff2673c637ff167c9f7d14691e1b98770f28b55467cc818a",
+      "size": 6190
+    },
+    "results/20260916T183751Z-verify2b/data_filter.json": {
+      "sha256": "e60da06955fe9e8a70a3de1ddaeb261e53af8f3a5e1d13ee62932d3773a1be24",
+      "size": 1572
+    },
+    "results/20260916T183751Z-verify2b/http_response.json": {
+      "sha256": "b8715eaf3ec74bd24569243798586f16a6ef7dffe330e98a54f22f094fe58735",
+      "size": 844
+    },
+    "results/20260916T183751Z-verify2b/manifest.json": {
+      "sha256": "e054c24da609c97b50fd3ecce33637bcc64f51c42ad723149bbbc2da786eeb70",
+      "size": 265
+    },
+    "results/20260916T183751Z-verify2b/reload_predictions.json": {
+      "sha256": "4e1ffbed45af3820bcf8c10d7638c925178e45284e9fc3cd86359ee230b80a67",
+      "size": 10453
+    },
+    "results/20260916T183751Z-verify2b/stress.json": {
+      "sha256": "efb4c0bc18cfca61511c7026a5b12e39fa423c00f0bfe63dd77343d03a52cbe5",
+      "size": 1163
+    },
+    "results/20260916T183751Z-verify2b/verification.json": {
+      "sha256": "49d95feaf1c72c076027b0aac71af26c80ecb6c7446755f88bd9171b79e6230c",
+      "size": 1920
+    },
+    "results/20260916T183823Z-train/correctness_final.json": {
+      "sha256": "5d6334a0c66ed567fa9709df4ea5b6040c88bb5258e864332e88be2bb9586634",
+      "size": 355
+    },
+    "results/20260916T183823Z-train/correctness_initial.json": {
+      "sha256": "6098996918d3fba844ad759459abb8063c2b7fdf9c65a891d7821a1d50dd9866",
+      "size": 423
+    },
+    "results/20260916T183823Z-train/data_filter.json": {
+      "sha256": "3a514e34a8a6f7e35776bafe5b19015d6c5cec0eaabea5558aa896471aba3b56",
+      "size": 2360
+    },
+    "results/20260916T183823Z-train/initial_validation_predictions.json": {
+      "sha256": "1189bd4f4d964b97e6fbeb8a10f8eaeb2c26e314e50b0a319b7ece322b325810",
+      "size": 324764
+    },
+    "results/20260916T183823Z-train/initial_validation_temperature_diagnostic.json": {
+      "sha256": "c466c9cb9b02206c5d1dab80db4fa579805f1ef890f549e6c27758240c35deda",
+      "size": 340
+    },
+    "results/20260916T183823Z-train/manifest.json": {
+      "sha256": "2cd93249a3c129203ddf3bb75827f810b30a61863e4dbf1bd5d2a1d7094419b8",
+      "size": 11460
+    },
+    "results/20260916T183823Z-train/summary.json": {
+      "sha256": "75044924b3f3b5a9b777627bccb1d80c7565b76a4520cc9c04228b0c7be93560",
+      "size": 11209
+    },
+    "results/20260916T183823Z-train/training.jsonl": {
+      "sha256": "2d425555511db013cae9ecafb28e2e1c7364c4dfcecd02b187f9c8c561a11e9e",
+      "size": 10422
+    },
+    "results/20260916T183823Z-train/validation.jsonl": {
+      "sha256": "964474c2ec7d6ee9c0f66055f51322944cc75f73a34adb5293849a3be6c1586c",
+      "size": 6173
+    },
+    "results/20260916T183823Z-train/validation_step_000040_predictions.json": {
+      "sha256": "e671e1508185765552b0f933ba03f356be62143c531d8ef534457d34b1645c9b",
+      "size": 333136
+    },
+    "results/20260916T185718Z-verify4b/data_filter.json": {
+      "sha256": "3a514e34a8a6f7e35776bafe5b19015d6c5cec0eaabea5558aa896471aba3b56",
+      "size": 2360
+    },
+    "results/20260916T185718Z-verify4b/http_255_choices_response.json": {
+      "sha256": "965b21465ccebb55e8c1b468da6e57653fe55dab5f2b48dec4ad207fdc322bfb",
+      "size": 12061
+    },
+    "results/20260916T185718Z-verify4b/http_long_context_response.json": {
+      "sha256": "fde2f6a2b43c93f77b292a5b89899254a7eee882598056a2a318d722c2b66f61",
+      "size": 216
+    },
+    "results/20260916T185718Z-verify4b/http_response.json": {
+      "sha256": "21ea4ea6c3e2c1c9e4383f5c5e340aaf6ee01b0903a9ce4d24672df63cf674d9",
+      "size": 861
+    },
+    "results/20260916T185718Z-verify4b/manifest.json": {
+      "sha256": "1fea5e941a3e38e11dc4e071c1dacb5de7c25345032207a95bb351e6cc211351",
+      "size": 265
+    },
+    "results/20260916T185718Z-verify4b/reload_predictions.json": {
+      "sha256": "2fee3e52111c3cd92babe0e36e5f2add009a4f69a3272a5fbacf7577026bc7f5",
+      "size": 10470
+    },
+    "results/20260916T185718Z-verify4b/stress.json": {
+      "sha256": "d93f33a15430685cc6357c9d2bdf08a10fa1e24d69d47fd58f66a4a9be7e46be",
+      "size": 1172
+    },
+    "results/20260916T185718Z-verify4b/verification.json": {
+      "sha256": "aede2c98c9773fd022e1ab807dfc66066a198dd01a9f35a6ca387e8e8341e48d",
+      "size": 2304
+    },
+    "results/20260916T185910Z-24h-launch/plan.json": {
+      "sha256": "979c1f0e0d4701c66115c209a25cf115d3210603a00d00ccd8b5ad027524b4fb",
+      "size": 649
+    },
+    "results/20260916T185910Z-24h-launch/resume_verification.json": {
+      "sha256": "772ab09829d7fb4403dcd7d3bb3685dee9e8559504b464191ff92f59a1a547d4",
+      "size": 933
+    },
+    "results/20260916T185910Z-24h-launch/state_snapshot.json": {
+      "sha256": "24cc3d6e9e17be546e68aea7654b4244d5b6bd883f5cf99ddb9e0c67cf87d7fa",
+      "size": 1328
+    },
+    "results/20260916T185910Z-24h-launch/training_correctness_initial.json": {
+      "sha256": "5d6334a0c66ed567fa9709df4ea5b6040c88bb5258e864332e88be2bb9586634",
+      "size": 355
+    },
+    "results/20260916T185910Z-24h-launch/training_manifest.json": {
+      "sha256": "b7c2e6e05d86f9c01a2aebd573e07ccd39da74bf25fd11e09d1150dc3326a619",
+      "size": 11618
+    },
+    "results/20260917-fleet-progress/ensemble-reference.json": {
+      "sha256": "3db8094bb4e01c2dfe74e880754b82c3ab08800529e357bdb4f1678beb21d40a",
+      "size": 6229
+    },
+    "results/20260917-fleet-progress/fixed-ensemble-validation.json": {
+      "sha256": "6c5dfa9d3528d4357cc89e90d10c7711b72eae50ddd67d67e15c4b902612c46d",
+      "size": 11460
+    },
+    "results/20260917-fleet-progress/fleet-four-candidates-status.json": {
+      "sha256": "2e2f2e6b8710562944b32c6d9b2619b415a6b79fd636dd85f1c1d1b48be186f4",
+      "size": 3234
+    },
+    "results/20260917-fleet-progress/four-candidate-registration.json": {
+      "sha256": "5d907228c8bc6ac48123359310bd5f56d567f2e72c30df3fd57d132c326f602d",
+      "size": 3568
+    },
+    "results/20260917-fleet-progress/gx10-status.json": {
+      "sha256": "3fb4d4c8d97d49d3f43460ff287bf8a619f863d4b93a9249b23b26f448b880a0",
+      "size": 4242
+    },
+    "results/20260917-fleet-progress/hf-final-watcher-launch.json": {
+      "sha256": "9654f1c752208de6c831fb7d6d9f1c0ec51c21f91351a148f9febd51aa579c41",
+      "size": 2519
+    },
+    "results/20260917-fleet-progress/hf-final-watcher-relaunch.json": {
+      "sha256": "a4e903e9eafeb4a911b0a532f8670b3b2b4f0cd9d324756009c078552a9c49d0",
+      "size": 1732
+    },
+    "results/20260917-fleet-progress/hf-initial-artifacts-publication.json": {
+      "sha256": "c8e938d34a13f17d3073c6ac4cb5de7b46f4cadf1e9f93f0e724766bf4f33e45",
+      "size": 2261
+    },
+    "results/20260917-fleet-progress/hf-snapshot-publication.json": {
+      "sha256": "686312f36485ada49c37bd2d1c129c39cdf86ff70316a87d65c0b0ef0cc363fc",
+      "size": 1031
+    },
+    "results/20260917-fleet-progress/hf-write-auth-verified.json": {
+      "sha256": "9e32210dd16078cf29f1339915f0712ba78925aadbe119d5a4c55b4f5512173f",
+      "size": 316
+    },
+    "results/20260917-fleet-progress/refinement-parent.json": {
+      "sha256": "f7c2765a5b9cf6fc794a30ec3b50bec986a046ed4deef080e3728b185c95a6ab",
+      "size": 1417
+    },
+    "results/20260917-fleet-progress/spark-a-status-20260917T021317Z.json": {
+      "sha256": "3c268f0e52bd5eb030f5d2799d51f6b33f52f750e7f277441c03ec83c24f93a4",
+      "size": 11252
+    },
+    "results/20260917-fleet-progress/spark-b-2b-completed-audit.json": {
+      "sha256": "53dea13fc44a987a071caa489ad2e832b0e434c516ada6de82423c5d2555aaef",
+      "size": 482109
+    },
+    "results/20260917-fleet-progress/spark-b-2b-completed-summary.json": {
+      "sha256": "16c665afdd0e3069ad2e18156da09c6dec8f58856b1ace9828feb559fd0e4620",
+      "size": 8395
+    },
+    "results/20260917-fleet-progress/spark-b-evidence-index.json": {
+      "sha256": "db281c7bfe8f023ed019038ab7c3e754fbdfda348a3eb5b2e9e4d90c73831b7c",
+      "size": 6475
+    },
+    "results/20260917-fleet-progress/spark-b-readonly-summary.json": {
+      "sha256": "16c665afdd0e3069ad2e18156da09c6dec8f58856b1ace9828feb559fd0e4620",
+      "size": 8395
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-best-validation-selection.json": {
+      "sha256": "a458754f2b605558fecc8c6506349d9cdb1b7d6bd25fb4f701d481b1e0a778bb",
+      "size": 19698
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-campaign-launch.json": {
+      "sha256": "203b65133d572f7b37bccfb6a4564ba5462efca651a797c3232d8c4420fede46",
+      "size": 4381
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-correctness-initial.json": {
+      "sha256": "f9ad8b0308a09577abc390412f214cc6289b1108ced3907570c6d4cf58539158",
+      "size": 354
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-current-status.json": {
+      "sha256": "a38095e1f9512022aad934a8db31d78985bfe0e336e4245c72203b6a5f68b7b2",
+      "size": 4735
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-data-filter.json": {
+      "sha256": "3a514e34a8a6f7e35776bafe5b19015d6c5cec0eaabea5558aa896471aba3b56",
+      "size": 2360
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-http-255-choices-response.json": {
+      "sha256": "b20caf2fb538c935bc5936c92c472082af58c54e4acd16dd2dc1447fe0b7d91f",
+      "size": 11995
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-http-long-context-response.json": {
+      "sha256": "8d13e7a618fa1a77a1cafead1afc6eee038411b3e09ff49d3e22afcbf6c0aa23",
+      "size": 215
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-http-response.json": {
+      "sha256": "cc339ccbd358dd410b52eaa9cb205971e8159eb3d2794ec2ee1a98f749943408",
+      "size": 864
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-inherited-validation-selection.json": {
+      "sha256": "a458754f2b605558fecc8c6506349d9cdb1b7d6bd25fb4f701d481b1e0a778bb",
+      "size": 19698
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-initial-predictions-verified.json": {
+      "sha256": "7817371003ac0883e10d6847fcd094cbc0f0ba9831ceedbb149e79aec9a23ccd",
+      "size": 707
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-initial-validation-selection.json": {
+      "sha256": "9675dfe39457a171238d81083d2c1e61dcb5aefa2429e2f2c9567ab79047e6c3",
+      "size": 19695
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-inputs-verified.json": {
+      "sha256": "8c5be28c56f988321f89dce076a1f6a834547795b2672b25d2d11ec60ad08bb4",
+      "size": 3472
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-pilot-correctness-final.json": {
+      "sha256": "f9ad8b0308a09577abc390412f214cc6289b1108ced3907570c6d4cf58539158",
+      "size": 354
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-pilot-correctness-initial.json": {
+      "sha256": "fd1b0d769f7a3f1cddacf00c19b0e3cec76c97c1d1e83a434ec2950bf178c677",
+      "size": 355
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-pilot-launch.json": {
+      "sha256": "d85c147ec85752b94eddc95b82ebdc6b93024d0915bdc0723ee65e941fde207d",
+      "size": 1419
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-pilot-manifest.json": {
+      "sha256": "e1ac6aaf86db2fff6c007a1018503877976ba04de7415125594142158b6f11a5",
+      "size": 12394
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-pilot-summary.json": {
+      "sha256": "929d4851d5b829a4cd69d337ab858f4f78480d6a654968d83141583c846fcec4",
+      "size": 11392
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-pilot-training.jsonl": {
+      "sha256": "6865669d51de01930927a807d94d34b78499004fbf6dfd78cba55ab26554ecb7",
+      "size": 2114
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-pilot-validation.jsonl": {
+      "sha256": "704b478c575080d30e1e555303b96b9307212d91c071673137f602a5b1db9fb6",
+      "size": 21145
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-pilot-verified.json": {
+      "sha256": "f39eebaabb44ff03e02353c93acdf7dafe512747b3778d4b4c553a7ceb7b78d4",
+      "size": 15161
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-plan.json": {
+      "sha256": "3e146d2daa7fa8e12689bebcfc8f3c8e27942328e15c2cf978a1950db21bb086",
+      "size": 777
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-reload-predictions.json": {
+      "sha256": "3c61f093082b06bf5274036d6e5f428465664ec9ae8125afc9ff99c781fc498a",
+      "size": 10396
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-resume-verified.json": {
+      "sha256": "0a2f072c250d91db2f419bd95c0c9eaee113229d3598f655d61ff05524773fcf",
+      "size": 3785
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-setup-status.json": {
+      "sha256": "c60e76f3a6a71b8893d1223db7091be667b3a95e02d9c919011ce4873b4b66e9",
+      "size": 199
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-startup-verified.json": {
+      "sha256": "97d75c66c1c1d76843cb2cb3a326780c452305501211da3b2ab3abef5a537b1b",
+      "size": 52576
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-stress.json": {
+      "sha256": "edc899aa2db8fae51a01818f0655a78b359b83dc521b3f7f399e458ffa5eeec7",
+      "size": 1166
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-training-manifest.json": {
+      "sha256": "0fe2c1bde4a47879d3ba040b97aa26c3025101dec10192e86188fe2cd98c33cf",
+      "size": 13002
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-verification-launch.json": {
+      "sha256": "ff87b649b1d2a9bb13d27e0bc8456d3fd5cd4076719e6833121ff69ca0775706",
+      "size": 1977
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-verification-manifest.json": {
+      "sha256": "997bee30fd9e1ea8420fb6671687b2e37c5aed76e6b1594521749bf73f35a01b",
+      "size": 264
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-verification.json": {
+      "sha256": "5d1a8d155bb53e27d2ffbe087d8b2e31686503fa277588c889d04865be3869c0",
+      "size": 2298
+    },
+    "results/20260917-fleet-progress/spark-b-refinement-warmstart-verified.json": {
+      "sha256": "72d8df546f492770738205561ecdc6c8c063b02c60f551ab752f3aafdebd78a7",
+      "size": 998
+    },
+    "results/20260917-fleet-progress/validation-trends.json": {
+      "sha256": "ddecc44c77276d0a6ca4d2418ad6f08a2e8caf00b327422e93f3407343610ac9",
+      "size": 4517
+    },
+    "results/20260917-playground/browser-verification.json": {
+      "sha256": "2806c5e75e560df7c0266e89aa92ac216bbca419be2094dee09e205d797ee460",
+      "size": 3596
+    },
+    "results/20260917-playground/concurrent-training.json": {
+      "sha256": "e581deac6f588ddf0a6effc595857a11bcbbf51979a5fdf812565e7611d608bc",
+      "size": 11763
+    },
+    "results/20260917-playground/desktop.png": {
+      "sha256": "0cbb5a6c9a9576ddceb95148024060350c5294779ed55063fb844f5978a3bf53",
+      "size": 138954
+    },
+    "results/20260917-playground/frontend-fixture-check.json": {
+      "sha256": "d368e15625343d2e94e3268a19e21fbd5f4a1cbe4994a4ef15b0542a83cfa1a6",
+      "size": 1179
+    },
+    "results/20260917-playground/launch.json": {
+      "sha256": "6af85b4328f8a253b3cc467290cad15eb0883268276eddeede0aa64d99395a54",
+      "size": 1991
+    },
+    "results/20260917-playground/mobile.png": {
+      "sha256": "e4711a2fa70b2d952e5f1b9e825dee5526c7eb6f837f99bf7f676aecd24c9af5",
+      "size": 127118
+    },
+    "results/20260917-playground/models.json": {
+      "sha256": "7a4e33a07801ab5fe918bb5a94c032900bd798513f6ab99608c9f06bb8c0572c",
+      "size": 1251
+    },
+    "results/20260917-playground/snapshots.json": {
+      "sha256": "a8187b1e1258b3c1ad50767c3e6b24d8e33b9533284f50347c8719c93ab1aea4",
+      "size": 1386
+    },
+    "results/checkpoint-sha256.txt": {
+      "sha256": "18fc16e95c38a28c1ec832f956c7ffc03fe9d0fdeb90348a84b73f2d9199937a",
+      "size": 254
+    },
+    "results/public-decisions-v1-manifest.json": {
+      "sha256": "adf5a8ca2bab60cf7429a82b1b7a2d1ae7d3cc812de04dd195a478dfd7c6f628",
+      "size": 5266
+    },
+    "scripts/campaign_status.py": {
+      "sha256": "1500b5e24f06231c7aafd7277aefd840582e35997e265f79db93614828d34411",
+      "size": 2301
+    },
+    "scripts/diagnose_parity.py": {
+      "sha256": "08b5d66d316ebda98a2226251a4f952701f86a1d5726ce7a4d7e8fb22755da5a",
+      "size": 3511
+    },
+    "scripts/download_candidate.py": {
+      "sha256": "d06a2c01be0cf6577f927fb37e3bc1eab014949fd934e4f4d9825674adba608e",
+      "size": 1141
+    },
+    "scripts/download_model.py": {
+      "sha256": "72ad9a5de44d09e2ee4ed8afb7c3c0ff6fb48987410a3f7f0368353572bf1f4d",
+      "size": 1013
+    },
+    "scripts/fleet_campaign.py": {
+      "sha256": "6f78b3df9ee3854330792d8218f91a4e907657e47bcca7f5d5e908044cbc5e78",
+      "size": 31181
+    },
+    "scripts/fleet_status.py": {
+      "sha256": "2519ced157ef4ac4fa449eebabae5740d7527778d578b4ac6720583010fa5217",
+      "size": 10289
+    },
+    "scripts/investigate_precision.py": {
+      "sha256": "609b744a926d8a45b87ba8d225e5312ee0c71a7096b21bd3589c5846e1dc847c",
+      "size": 10519
+    },
+    "scripts/launch_24h.py": {
+      "sha256": "39c26dc10535d3adac09209b0743edf2ed384563e732512ad62f8c6f37161a83",
+      "size": 19127
+    },
+    "scripts/prepare_public_data.py": {
+      "sha256": "32ea84aa719818e1141b958b6ef27a85f7ddb86bfcd7c1585fc487b25253d977",
+      "size": 11886
+    },
+    "scripts/publish_hf_final.py": {
+      "sha256": "278efc5878d7ebc5d1171f3a735d6c7d78275d93650e5a8a16c3f55b8353a577",
+      "size": 20356
+    },
+    "scripts/publish_hf_snapshot.py": {
+      "sha256": "b5fe16a00fcbc5ab97121428c6ce750275ee193438e3ffe24aac5e3bb325c018",
+      "size": 15165
+    },
+    "scripts/run_experiment.sh": {
+      "sha256": "661a6309fc54a2a8aff918f14a553c72dcb21730bd6a3cfc55d6ccd4700d11d3",
+      "size": 1213
+    },
+    "scripts/run_precision.sh": {
+      "sha256": "766d82b30cf3e83951f662685b4472ee053c7fbc171c4ff140823bf4d1c2782f",
+      "size": 998
+    },
+    "scripts/run_smoke.sh": {
+      "sha256": "39d59f2120f362729d1c2e384391b82be1e580dcc1aca0eaa7ab231115225574",
+      "size": 1054
+    },
+    "scripts/start_spark_candidate.sh": {
+      "sha256": "c2ca18b008a144de7cb264c9fcca634d68c4a8a317db3638e8a70b3dbcff064b",
+      "size": 6545
+    },
+    "scripts/verify_artifact.py": {
+      "sha256": "9833350e9d72c0065b15206bb71c5a8b5a6b3185219db90074e369ede563a985",
+      "size": 9971
+    },
+    "scripts/verify_playground.cjs": {
+      "sha256": "0be5180bf1d2f812a7e416099b3b6263ef3fa766b2b1057418495eb9592fa48e",
+      "size": 5866
+    },
+    "selection.py": {
+      "sha256": "be0a7a8496b5b830aa572ceba93606320f442063fd38180503fd6980dc1c578f",
+      "size": 5335
+    },
+    "smoke_data.py": {
+      "sha256": "06b3cbac1c8c4a86b8aecbee4459073cc3e46d4ddcd576392f3cb4805924f815",
+      "size": 1807
+    },
+    "smoke_train.py": {
+      "sha256": "8cdeb2b397177fc9c26638aaa871501ddab8e3871aa1573ecd98f66590f5c228",
+      "size": 19630
+    },
+    "tests/test_campaign.py": {
+      "sha256": "90b132f655db9e9fd7b71c4916d47064c51e65cc1cc0de1a69f0b0cdc8d1cad7",
+      "size": 8567
+    },
+    "tests/test_fleet_campaign.py": {
+      "sha256": "ed5fc03bdc5d5c3de28b1069aadb93e9ca1d30965e412bcafc43dd97d86724e3",
+      "size": 15318
+    },
+    "tests/test_fleet_status.py": {
+      "sha256": "681eabe15b1de8380302a15f87cba411037da44c73a5e8b817455a7bb13dd061",
+      "size": 3469
+    },
+    "tests/test_playground.py": {
+      "sha256": "f63b58eef3b0f9c42aa3d445ceb4fa5462e936deb7b11e2e0fc962af6ed32c96",
+      "size": 10663
+    },
+    "tests/test_publish_hf_final.py": {
+      "sha256": "dd18fb43526d23e36166787a1d3eae5a12aedd6ffc54951459c97c37b1f2c3ee",
+      "size": 15397
+    },
+    "tests/test_publish_hf_snapshot.py": {
+      "sha256": "f48905828b51060f9a265505245eb10f814b47b8610ce5b01d67b2005c828edd",
+      "size": 9332
+    },
+    "tests/test_scorer.py": {
+      "sha256": "2c6f5d5e9ff634049cbe9c88b5f126710298b65ce603d7ced2072bc06e3974d8",
+      "size": 5091
+    },
+    "tests/test_selection.py": {
+      "sha256": "09fc10fd394ea870307175698a81b516a0bafc536757be918f5cacb86d124770",
+      "size": 4090
+    },
+    "tests/test_training_harness.py": {
+      "sha256": "df8175b7d051625142844cece4b1bf709f395a39f64319c8eeb4f6edd8140b87",
+      "size": 22141
+    },
+    "training_model.py": {
+      "sha256": "d5b0aefeeb5290816bc0b669aa0a8cbbe27f6a12b9cb23c141ac9b9ae9ee4e65",
+      "size": 7862
+    },
+    "web/playground.css": {
+      "sha256": "09969ed5332298048595879c930d1065850db3c915688ea2da343845e879f27b",
+      "size": 14088
+    },
+    "web/playground.html": {
+      "sha256": "69495feec84f29a6efd06c7e45c1ddac3d37717a16b7b9036093609dfd24b4df",
+      "size": 8847
+    },
+    "web/playground.js": {
+      "sha256": "b2ccc2dab0fb670fb3cd11e43f641cdab7c3752724351e7641f4dd3a4ca86646",
+      "size": 19280
+    }
+  },
+  "source_commit": "b05f7e3975a974c26ab667c6fdf91ec489a8cec0"
+}

sources/b05f7e3975a974c26ab667c6fdf91ec489a8cec0/source.tar.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:64a88e5e218f68e4b1dda00e7012eee0aaba4e4fcb9e0443b081c188528946c9
+size 1654578