Model card: archive public, resume receipt 15:41 UTC (47accac8)

EVALUATION.json

@@ -1,8 +1,8 @@
 {
   "schema_version": 1,
   "measurement_owner": "SlayerLab / Fabryka AI",
-  "status": "descriptive; training-lineage contamination scan pending",
-  "source_aggregate_sha256": "2d257a95523605b06aa964cb974f65d86ab025875cc1f87dc98b242b90729440",
+  "status": "descriptive intermediate base checkpoint measurements; at the 12 November 2026 14:53 UTC archive snapshot the separate repository was public and native .pt upload was in progress; verify individual file availability and checksums in that repository; main repository contains documentation; lineage checks incomplete",
+  "source_aggregate_sha256": "d91456c7480f8031288afd5056c002ce68ef7d7751556d222ae54d7ab6fc648a",
   "source_generator_sha256": null,
   "benchmark_community_report": "https://github.com/slayerlabs/benchmarks-fabryka-ai/blob/8b547007526971bd2fc35dfe56ff2c7d9b802568/research/results/full-run-2026-11/README.md",
   "community_report_is_separate_measurement": true,
@@ -38,6 +38,25 @@
       "belebele_pol": "lm-eval belebele_pol_Latn 0-shot: ctx='P: {flores_passage}\\nQ: {question.strip()}\\nA: {mc_answer1}\\nB: ...\\nC: ...\\nD: {mc_answer4}\\nAnswer:', kont=' A'..' D', gold=int(correct_answer_num)-1; 0-shot; LL = suma log p tokenów kontynuacji; start id 0; add_special_tokens=False; _encode_pair jak HFLM; obcięcie z lewej do block; fp32",
       "hellaswag_pl": "OpenGPTX ogx_hellaswagx_pl @0018eb9b (= lm-eval hellaswag process_docs na polskich polach); 0-shot; LL = suma log p tokenów kontynuacji; start id 0; add_special_tokens=False; _encode_pair jak HFLM; obcięcie z lewej do block; fp32",
       "mmlu_pl": "OpenGPTX ogx_mmlux_pl-* @0018eb9b 0-shot: ctx=description(PL, doklejone BEZ separatora, jak w forku)+'{question.strip()}\\nA. {c0}\\nB. {c1}\\nC. {c2}\\nD. {c3}\\nOdpowiedź:', kont=' A'..' D', gold=answer; acc = mikro po 14042 pozycjach (jak grupa lm-eval mmlu, wagi = rozmiary); 0-shot; LL = suma log p tokenów kontynuacji; start id 0; add_special_tokens=False; _encode_pair jak HFLM; obcięcie z lewej do block; fp32"
+    },
+    "reference_baselines": {
+      "uniform_two_options": {
+        "accuracy": 0.5,
+        "assumption": "uniform independent choice between two options"
+      },
+      "uniform_four_options": {
+        "accuracy": 0.25,
+        "assumption": "uniform independent choice among four options; reference context, not an empirical model"
+      },
+      "boolq_always_yes": {
+        "accuracy": 0.6217125382262997,
+        "correct": 2033,
+        "n": 3270,
+        "dataset_repo": "aps/super_glue",
+        "dataset_sha": "3de24cf8022e94f4ee4b9d55a6f539891524d646",
+        "split": "boolq validation",
+        "provenance": "aggregate class counts supplied by measurement reviewer; not per-item labels"
+      }
     }
   },
   "ci": {
@@ -2081,6 +2100,260 @@
           "correct_primary": 3470
         }
       }
+    },
+    "1.09-k35000": {
+      "step": 35000,
+      "weights_sha256": "c1d53f72bfaaa68db13fa2b6ada34ccd29a3aa7b7161bbf8733e13feb88e15b5",
+      "source_results_sha256": "01885b60efca4b0b1605502e5a1d9f02af7ed1718c4255c8d7c9a0288cac36c0",
+      "evaluator_sha256": "ae6658f9c83a484e373dc6d14476a7aeb98a905c27cd6d97ea234fb56140e016",
+      "weights_published": false,
+      "tasks": {
+        "arc_challenge": {
+          "n": 1172,
+          "n_pelne": 1172,
+          "dataset_repo": "allenai/ai2_arc",
+          "dataset_sha": "210d026faf9955653af8916fad021475a3f00453",
+          "pliki_sha256": {
+            "ARC-Challenge/test-00000-of-00001.parquet": "62f03257e737aed263f55c6abf87c7bb0028a44a6bdd2a26eb1279eb42c1d1e9"
+          },
+          "acc": 0.33959,
+          "acc_ci95": [
+            0.31314,
+            0.366894
+          ],
+          "acc_norm": 0.351536,
+          "acc_norm_ci95": [
+            0.324232,
+            0.37884
+          ],
+          "split": "test",
+          "correct_primary": 412
+        },
+        "hellaswag": {
+          "n": 10042,
+          "n_pelne": 10042,
+          "dataset_repo": "Rowan/hellaswag",
+          "dataset_sha": "218ec52e09a7e7462a5400043bb9a69a41d06b76",
+          "pliki_sha256": {
+            "data/validation-00000-of-00001.parquet": "899813071e1e95efafec90f856e1987d2150fa4d020fc005df6962c259f660cd"
+          },
+          "acc": 0.409978,
+          "acc_ci95": [
+            0.400219,
+            0.419638
+          ],
+          "acc_norm": 0.524796,
+          "acc_norm_ci95": [
+            0.515037,
+            0.534555
+          ],
+          "split": "validation",
+          "correct_primary": 5270
+        },
+        "piqa": {
+          "n": 1838,
+          "n_pelne": 1838,
+          "dataset_repo": "baber/piqa",
+          "dataset_sha": "142f6d7367fd9877f0fb3b5734ea6a545f54cdd1",
+          "pliki_sha256": {
+            "piqa_validation.parquet": "2b17ee1215b1ceb2ddabf4c90a05c5a8b833bef9d207960f57d111c43a2080f3"
+          },
+          "acc": 0.708923,
+          "acc_ci95": [
+            0.688248,
+            0.729611
+          ],
+          "acc_norm": 0.712731,
+          "acc_norm_ci95": [
+            0.691513,
+            0.732862
+          ],
+          "split": "validation",
+          "correct_primary": 1310
+        },
+        "winogrande": {
+          "n": 1267,
+          "n_pelne": 1267,
+          "dataset_repo": "allenai/winogrande",
+          "dataset_sha": "01e74176c63542e6b0bcb004dcdea22d94fb67b5",
+          "pliki_sha256": {
+            "winogrande_xl/validation-00000-of-00001.parquet": "f9d914d1818c0aba98cabdf8af2bc4bd943c462b9c5e7062647e05c2ce3315d1"
+          },
+          "acc": 0.567482,
+          "acc_ci95": [
+            0.540647,
+            0.594317
+          ],
+          "acc_norm": 0.567482,
+          "acc_norm_ci95": [
+            0.539858,
+            0.595107
+          ],
+          "split": "validation",
+          "correct_primary": 719
+        },
+        "boolq": {
+          "n": 3270,
+          "n_pelne": 3270,
+          "dataset_repo": "aps/super_glue",
+          "dataset_sha": "3de24cf8022e94f4ee4b9d55a6f539891524d646",
+          "pliki_sha256": {
+            "boolq/validation-00000-of-00001.parquet": "f46438b8cce048184552a2bfcdb8252eadb3a2602f0c22334caccd9b6ea547d6"
+          },
+          "acc": 0.598777,
+          "acc_ci95": [
+            0.581957,
+            0.615596
+          ],
+          "acc_norm": 0.621407,
+          "acc_norm_ci95": [
+            0.604893,
+            0.63792
+          ],
+          "split": "validation",
+          "correct_primary": 1958
+        },
+        "openbookqa": {
+          "n": 500,
+          "n_pelne": 500,
+          "dataset_repo": "allenai/openbookqa",
+          "dataset_sha": "388097ea7776314e93a529163e0fea805b8a6454",
+          "pliki_sha256": {
+            "main/test-00000-of-00001.parquet": "cd5483e366daa230c1c87bbdc512d8b7229f14f6dd04d19fc8b1a3855aaaa8a3"
+          },
+          "acc": 0.244,
+          "acc_ci95": [
+            0.20795,
+            0.282
+          ],
+          "acc_norm": 0.344,
+          "acc_norm_ci95": [
+            0.304,
+            0.386
+          ],
+          "split": "test",
+          "correct_primary": 172
+        },
+        "belebele_pol": {
+          "n": 900,
+          "n_pelne": 900,
+          "dataset_repo": "facebook/belebele",
+          "dataset_sha": "7899cdfa4e1e0d733fd77c848e2c273cb1d32be2",
+          "pliki_sha256": {
+            "data/pol_Latn.jsonl": "1699bb60b65e258d71989c541027c5439eb36a4f5032a5be392137bf5a1b6bec"
+          },
+          "acc": 0.243333,
+          "acc_ci95": [
+            0.215556,
+            0.271139
+          ],
+          "acc_norm": 0.243333,
+          "acc_norm_ci95": [
+            0.215556,
+            0.271111
+          ],
+          "split": "pol_Latn evaluation set",
+          "correct_primary": 219
+        },
+        "hellaswag_pl": {
+          "n": 10042,
+          "n_pelne": 10042,
+          "dataset_repo": "Eurolingua/hellaswagx",
+          "dataset_sha": "c38911a204759866dbcde35eeb584c36caf11a8b",
+          "pliki_sha256": {
+            "hellaswag_PL_validation.jsonl": "acdfc6d397885a4c86de178a6e5f9b7e9b90a7ff81ffcd6e5e0d3d0788dd3a16"
+          },
+          "acc": 0.362079,
+          "acc_ci95": [
+            0.352619,
+            0.37144
+          ],
+          "acc_norm": 0.466341,
+          "acc_norm_ci95": [
+            0.456779,
+            0.4763
+          ],
+          "split": "validation",
+          "correct_primary": 4683
+        },
+        "mmlu_pl": {
+          "n": 14042,
+          "n_pelne": 14042,
+          "dataset_repo": "Eurolingua/mmlux",
+          "dataset_sha": "7fb62cf05c1ce9ea550dbfd2ea6c145204344882",
+          "pliki_sha256": {
+            "hendrycks_abstract_algebra_PL_test.jsonl": "4905c6ef8d306c2ff1e1771ef8c15b63152bb88effec11844217893b81fcc3ea",
+            "hendrycks_anatomy_PL_test.jsonl": "5cba9bc54a6f0bdaddcd2e0e23155fccfeab469b561f03cc571e1387abeb9a38",
+            "hendrycks_astronomy_PL_test.jsonl": "00dc5fca7a8be3d8bef9e24ad9018e5417fb07703ebd2d0039c2c82418272b03",
+            "hendrycks_business_ethics_PL_test.jsonl": "9c9542880c720e2d0f5a768ab2d6602bdf5c394d66e51ff352af79aaca6bbd96",
+            "hendrycks_clinical_knowledge_PL_test.jsonl": "79fd5fd7a17dc7c7d84a9e9b138cf13c2dcaf6ab332eb1a0f299a7b52dc89c61",
+            "hendrycks_college_biology_PL_test.jsonl": "b49d766fb2c4160809d3907b846cd6c5c609299f47d9096bcf3208b843392c48",
+            "hendrycks_college_chemistry_PL_test.jsonl": "36c6fd391411789310dca0bbd36aec9aae8784df2bac2e6838e0895eebd97f1d",
+            "hendrycks_college_computer_science_PL_test.jsonl": "90ab29151dc4b8c7e13893516474201b73e8f9765cdf87225b6b9cacdd540e70",
+            "hendrycks_college_mathematics_PL_test.jsonl": "84411ee6464d01f24f2482bccb451635fe75d8169e33932ce6e067c1be3d37ee",
+            "hendrycks_college_medicine_PL_test.jsonl": "faa451d891c8a67b15bae4921c49a8afea448ac7b1b6095646c3570ac3040606",
+            "hendrycks_college_physics_PL_test.jsonl": "bfcaac4cb6d13e7cbcc6ebfb7ea6a9ab0d0ec07613b1b10ca679e632106b4bc2",
+            "hendrycks_computer_security_PL_test.jsonl": "c8324fb94b5b4b7e17440520ea20a22d8c03c3b86da81824059602982463cf8a",
+            "hendrycks_conceptual_physics_PL_test.jsonl": "5227e303a5437f437a4ad9d7eb75917d36051ce635af9f43a92d9525918eac64",
+            "hendrycks_econometrics_PL_test.jsonl": "9bf1d2a2c9eb4f5692b3876919dbc4e679d43d3ea2b64dc8fa2c17703a3e696b",
+            "hendrycks_electrical_engineering_PL_test.jsonl": "a431be1d342cc1eca6aa408688b78751c74621056a4fa4c31e7dfda443e4be6f",
+            "hendrycks_elementary_mathematics_PL_test.jsonl": "9cdb8692fc08d6d8ac5dec24d89f386863b6a02480eb3b8c5be641f15700e667",
+            "hendrycks_formal_logic_PL_test.jsonl": "0f30d48346a0cd2f136463b283d8bd3fe085de67da2a34d90cac689740f9b5be",
+            "hendrycks_global_facts_PL_test.jsonl": "21c753017550f0d31d7c2cdbe82637c877682a9edf8ab913de720e79c82b374f",
+            "hendrycks_high_school_biology_PL_test.jsonl": "7b086933dd97cfc392fba5f78c3523ddc0302cf2b7caed7428a75bfdf9fbb7f2",
+            "hendrycks_high_school_chemistry_PL_test.jsonl": "85f836040875313d60cf9241bd9e827a3cdb770cf159ddf3f7032edb293871ec",
+            "hendrycks_high_school_computer_science_PL_test.jsonl": "034b7b497ada5cc1146c6c5fa02291094d5da512bc0c618ae2b967bd941a70c1",
+            "hendrycks_high_school_european_history_PL_test.jsonl": "023ce8f773cb5dccd86658165a86a5bc1b0d83d53cede5a22c8f7a027582e2a4",
+            "hendrycks_high_school_geography_PL_test.jsonl": "3cce3d6defc5cad9742f9db4407ec5176e6b5d1634b0f8ecb4cfe9f4cabfbe5a",
+            "hendrycks_high_school_government_and_politics_PL_test.jsonl": "9c73c16602cd663acf3b08cdecef6aa8a2c019ed647f02cb970f078d88e3ec7d",
+            "hendrycks_high_school_macroeconomics_PL_test.jsonl": "9425c91f59bd35355262f0c60b04b479a2e700064751efddadb9a28cbc83974d",
+            "hendrycks_high_school_mathematics_PL_test.jsonl": "80c73fa2b882104047260829a57cdca588d1ec9ba690b8a8a5cb87f65283e735",
+            "hendrycks_high_school_microeconomics_PL_test.jsonl": "aebc508914a9b32f5dd1bdc3ebd1375a619791214cb67968ccc6d5712675cbe8",
+            "hendrycks_high_school_physics_PL_test.jsonl": "bba7751511cada0b7b0b07133f2a961e5f046e174e201628a6722a279ab4e47f",
+            "hendrycks_high_school_psychology_PL_test.jsonl": "adf5943b42493ae59f9310028a45adffaace2359d52b8cbb1fa9cc41d6371e45",
+            "hendrycks_high_school_statistics_PL_test.jsonl": "14ffdec0e35d7437d7b2243c3bcbf606843df9ac4eaa044cae95faa3f6e41fdf",
+            "hendrycks_high_school_us_history_PL_test.jsonl": "d8a7946dbf1860b1e99ff0d1948ede18500cc06984ef68786ffed6a8160731f7",
+            "hendrycks_high_school_world_history_PL_test.jsonl": "ae943acd33cf4b23fd3d70d4cc5b0d16f073574c547de91bfc157d552bf63ec0",
+            "hendrycks_human_aging_PL_test.jsonl": "21df1fd8d194a6e453e71e1583626fc723f1c9b44cc571ef7d6074450754ed82",
+            "hendrycks_human_sexuality_PL_test.jsonl": "d7c467fcee3671069986948b69e6ec69c2630173207025ce4925719bb1a880d4",
+            "hendrycks_international_law_PL_test.jsonl": "f0b6f86e54390b63fda0ecf07c7c55f87d30864f6a698c9819f8d8e6eca4d746",
+            "hendrycks_jurisprudence_PL_test.jsonl": "9d7f42f41510da5d835d8df834c4bf6bcd2e03e658cc906ec334810b30ebbe96",
+            "hendrycks_logical_fallacies_PL_test.jsonl": "8b7cd34e9126adcfbc4b9b7bee8345d97352ec892221a0e98fa268210b04ee51",
+            "hendrycks_machine_learning_PL_test.jsonl": "f654c60a51877a9d269471ab01b54b0d4d20f9cc07f7fb57512fb3e90d5a8608",
+            "hendrycks_management_PL_test.jsonl": "628374b8f9beabefe43c2b1dda2c07760eb455d2e15f9e58f58506d61f0fc3bb",
+            "hendrycks_marketing_PL_test.jsonl": "31dc1eef1527c98642080c0b3866ea113bda77c2a03c5155aafe382206ffacdb",
+            "hendrycks_medical_genetics_PL_test.jsonl": "f05fcc33d029460a7f7916ec83debfc746ced5ea931f7a56b63fe1245819e4a4",
+            "hendrycks_miscellaneous_PL_test.jsonl": "94be7854c05aef25ebca4ae1dc562f37172b5346a14fb760b8e89eefea588809",
+            "hendrycks_moral_disputes_PL_test.jsonl": "245ad1cf635274e5ac81559542cb21638b7c729e8a811f76271f4301937d6127",
+            "hendrycks_moral_scenarios_PL_test.jsonl": "d60b08bb71451594385db5ff6ffdef4ca11fec801dc86d681e68f1727a690f9b",
+            "hendrycks_nutrition_PL_test.jsonl": "3cfe618d8e23e3c8e5c752544f07a71b90b4c84d4bc5d673026e8be4de4bf285",
+            "hendrycks_philosophy_PL_test.jsonl": "a1939f9773085bdae7e127fce099afc4153ec4cbe5f66e126f12599feaf76639",
+            "hendrycks_prehistory_PL_test.jsonl": "42eaf871b859f22321dfdec21b6b6b4a3c8f1435c2d4cd2da46dc9fb5ceced88",
+            "hendrycks_professional_accounting_PL_test.jsonl": "1fb6440616a195b94d3de760af5cacd05656667dfac595e2601504cda407c25e",
+            "hendrycks_professional_law_PL_test.jsonl": "fb86b175ecdfe8b56d4991bf6d6705a58f1d89a3efda2a8eb5f01f8082b21e57",
+            "hendrycks_professional_medicine_PL_test.jsonl": "ec818eb13a66e631ab82a10bae1536377dcd4bcadf2321c964b7718c60992080",
+            "hendrycks_professional_psychology_PL_test.jsonl": "72ddab1aaddc26266f4964941334a84d9243f63c329d2d9349bdcaa20efa6bc3",
+            "hendrycks_public_relations_PL_test.jsonl": "2e16835d88ba1ebb6f70a526bd85b9830cb1d3e14e2358a43ce15a3d548232e1",
+            "hendrycks_security_studies_PL_test.jsonl": "c476a634f4f48ff025869158057de726a8080bee253d88921d90252204aa3e4c",
+            "hendrycks_sociology_PL_test.jsonl": "dd56860da95ed793943c9c201f9aa8f5aa8d83753b1bbcbc440fff7ac30f881a",
+            "hendrycks_us_foreign_policy_PL_test.jsonl": "95183b3fcd02d17f5bb7d359b709b3fde13bf6f0f0f1375ab676b0e7dda36786",
+            "hendrycks_virology_PL_test.jsonl": "99e2c6f80907315317619f8c92f8d501ae484619c1169e36d9ad8fa59abfbda7",
+            "hendrycks_world_religions_PL_test.jsonl": "4631dbc58d379cbc46751515f36af682f8bf53db36f3345dbe090d5a00b7b24a"
+          },
+          "acc": 0.232802,
+          "acc_ci95": [
+            0.225823,
+            0.239709
+          ],
+          "acc_norm": 0.232802,
+          "acc_norm_ci95": [
+            0.225823,
+            0.239923
+          ],
+          "split": "test; micro-average over subjects",
+          "correct_primary": 3269
+        }
+      }
     }
   },
   "additional_measurements": {
@@ -2280,5 +2553,475 @@
       }
     }
   },
-  "source_generator_note": "The updated approved aggregate does not supply its CI-generator SHA; evaluator SHA is retained per model. Previous aggregate: ce0822a324f805bb61f28cc8ea3197e6a9abd97a4591239349dc60141f5289db; previous CI generator: b792b5a7b4f5218acff4b3d5e9fb63c6f11d39fa26cca6fbd467758c90555a34."
+  "source_generator_note": "The current aggregate adds update 35000 and preserves all earlier scores, intervals, counts and dataset pins from aggregate 2d257a95523605b06aa964cb974f65d86ab025875cc1f87dc98b242b90729440. It records Binomial(n,p)/n percentile bootstrap, seed 20261110, but does not supply random-stream derivation or CI-generator SHA. Evaluator SHA is retained per model.",
+  "auxiliary_diagnostics": {
+    "source_aggregate_sha256": "c8ce6783e5d6fa2a814a878a881e14861d5303bc8c921503ced80b02c63d7db5",
+    "method": "unweighted mean of class recalls (balanced accuracy); collected after the primary measurements; not a replacement for primary accuracy",
+    "units": "percent",
+    "uncertainty": "confidence intervals not supplied for these auxiliary values; no formal checkpoint-difference conclusion",
+    "models": {
+      "1.09-k5000": {
+        "source_diagnostic_sha256": "118d60cb6018095ac1bd175b550e0dd3c443c98fa1be9acb869fed2d9de584ac",
+        "source_track_results_sha256": "ec8d493463d99687f85a06654cef7706b3d37f0c9b9c3e760512f2baac34d632",
+        "tasks": {
+          "boolq": {
+            "n": 3270,
+            "balanced_accuracy_percent": 53.39,
+            "classes": [
+              "no",
+              "yes"
+            ],
+            "predicted_class_percent": [
+              14.6,
+              85.4
+            ],
+            "gold_class_percent": [
+              37.8,
+              62.2
+            ]
+          },
+          "mmlu_pl": {
+            "n": 14042,
+            "balanced_accuracy_percent": 25.34,
+            "classes": [
+              "A",
+              "B",
+              "C",
+              "D"
+            ],
+            "predicted_class_percent": [
+              72.3,
+              6.8,
+              7.0,
+              13.9
+            ],
+            "gold_class_percent": [
+              22.9,
+              24.7,
+              25.5,
+              26.9
+            ]
+          },
+          "belebele_pol": {
+            "n": 900,
+            "balanced_accuracy_percent": 25.6,
+            "classes": [
+              "A",
+              "B",
+              "C",
+              "D"
+            ],
+            "predicted_class_percent": [
+              84.7,
+              5.0,
+              7.3,
+              3.0
+            ],
+            "gold_class_percent": [
+              22.9,
+              27.9,
+              27.3,
+              21.9
+            ]
+          }
+        }
+      },
+      "1.09-k10000": {
+        "source_diagnostic_sha256": "52b169909e2774cbaa7ed2183e644ea077e38c6f6fa941a348a7c1207c0c7c1d",
+        "source_track_results_sha256": "3fff987f3b9d74eaba1e121a6a769247fbc12c79043bd5cb5d442c8dafeb9a69",
+        "tasks": {
+          "boolq": {
+            "n": 3270,
+            "balanced_accuracy_percent": 52.42,
+            "classes": [
+              "no",
+              "yes"
+            ],
+            "predicted_class_percent": [
+              13.6,
+              86.4
+            ],
+            "gold_class_percent": [
+              37.8,
+              62.2
+            ]
+          },
+          "mmlu_pl": {
+            "n": 14042,
+            "balanced_accuracy_percent": 25.09,
+            "classes": [
+              "A",
+              "B",
+              "C",
+              "D"
+            ],
+            "predicted_class_percent": [
+              89.0,
+              4.4,
+              3.2,
+              3.4
+            ],
+            "gold_class_percent": [
+              22.9,
+              24.7,
+              25.5,
+              26.9
+            ]
+          },
+          "belebele_pol": {
+            "n": 900,
+            "balanced_accuracy_percent": 25.4,
+            "classes": [
+              "A",
+              "B",
+              "C",
+              "D"
+            ],
+            "predicted_class_percent": [
+              80.1,
+              6.8,
+              10.2,
+              2.9
+            ],
+            "gold_class_percent": [
+              22.9,
+              27.9,
+              27.3,
+              21.9
+            ]
+          }
+        }
+      },
+      "1.09-k15000": {
+        "source_diagnostic_sha256": "5fbf1fdc97b8d09e050c66c0385970140740ea6eccb84dde149d511eaecb88ea",
+        "source_track_results_sha256": "1bbb4ce856f45ed4ebacc1911431b749f6df4ed8f5caf292437c9424f03959bd",
+        "tasks": {
+          "boolq": {
+            "n": 3270,
+            "balanced_accuracy_percent": 53.98,
+            "classes": [
+              "no",
+              "yes"
+            ],
+            "predicted_class_percent": [
+              24.8,
+              75.2
+            ],
+            "gold_class_percent": [
+              37.8,
+              62.2
+            ]
+          },
+          "mmlu_pl": {
+            "n": 14042,
+            "balanced_accuracy_percent": 24.99,
+            "classes": [
+              "A",
+              "B",
+              "C",
+              "D"
+            ],
+            "predicted_class_percent": [
+              74.4,
+              17.1,
+              3.7,
+              4.7
+            ],
+            "gold_class_percent": [
+              22.9,
+              24.7,
+              25.5,
+              26.9
+            ]
+          },
+          "belebele_pol": {
+            "n": 900,
+            "balanced_accuracy_percent": 24.42,
+            "classes": [
+              "A",
+              "B",
+              "C",
+              "D"
+            ],
+            "predicted_class_percent": [
+              93.8,
+              0.2,
+              5.4,
+              0.6
+            ],
+            "gold_class_percent": [
+              22.9,
+              27.9,
+              27.3,
+              21.9
+            ]
+          }
+        }
+      },
+      "1.09-k20000": {
+        "source_diagnostic_sha256": "a75346af9bdbaa6773a0a76264f4f606d6cad1ef8e77351c2368fd977a0c3b3f",
+        "source_track_results_sha256": "a6225b2af295c846852314f88c0e08f3a71d67d3cd67a53c65bdd87b989a018f",
+        "tasks": {
+          "boolq": {
+            "n": 3270,
+            "balanced_accuracy_percent": 57.34,
+            "classes": [
+              "no",
+              "yes"
+            ],
+            "predicted_class_percent": [
+              72.6,
+              27.4
+            ],
+            "gold_class_percent": [
+              37.8,
+              62.2
+            ]
+          },
+          "mmlu_pl": {
+            "n": 14042,
+            "balanced_accuracy_percent": 25.06,
+            "classes": [
+              "A",
+              "B",
+              "C",
+              "D"
+            ],
+            "predicted_class_percent": [
+              61.4,
+              25.8,
+              5.5,
+              7.3
+            ],
+            "gold_class_percent": [
+              22.9,
+              24.7,
+              25.5,
+              26.9
+            ]
+          },
+          "belebele_pol": {
+            "n": 900,
+            "balanced_accuracy_percent": 24.91,
+            "classes": [
+              "A",
+              "B",
+              "C",
+              "D"
+            ],
+            "predicted_class_percent": [
+              57.6,
+              8.7,
+              33.6,
+              0.2
+            ],
+            "gold_class_percent": [
+              22.9,
+              27.9,
+              27.3,
+              21.9
+            ]
+          }
+        }
+      },
+      "1.09-k25000": {
+        "source_diagnostic_sha256": "5fe51476cb474495c37a5aab8ca50f1f78da8a6022595d4302ed95912b608511",
+        "source_track_results_sha256": "7b1ac4f7710b11b64542eb193a2820144c3ba6482901159b20356d2928e6f9c5",
+        "tasks": {
+          "boolq": {
+            "n": 3270,
+            "balanced_accuracy_percent": 57.44,
+            "classes": [
+              "no",
+              "yes"
+            ],
+            "predicted_class_percent": [
+              69.7,
+              30.3
+            ],
+            "gold_class_percent": [
+              37.8,
+              62.2
+            ]
+          },
+          "mmlu_pl": {
+            "n": 14042,
+            "balanced_accuracy_percent": 24.77,
+            "classes": [
+              "A",
+              "B",
+              "C",
+              "D"
+            ],
+            "predicted_class_percent": [
+              32.9,
+              44.6,
+              8.1,
+              14.4
+            ],
+            "gold_class_percent": [
+              22.9,
+              24.7,
+              25.5,
+              26.9
+            ]
+          },
+          "belebele_pol": {
+            "n": 900,
+            "balanced_accuracy_percent": 26.25,
+            "classes": [
+              "A",
+              "B",
+              "C",
+              "D"
+            ],
+            "predicted_class_percent": [
+              72.7,
+              9.9,
+              16.1,
+              1.3
+            ],
+            "gold_class_percent": [
+              22.9,
+              27.9,
+              27.3,
+              21.9
+            ]
+          }
+        }
+      },
+      "1.09-k30000": {
+        "source_diagnostic_sha256": "97c275abfea2688e29f2016dbb7551e15f5f2d55bd3453641bd8a2be73268446",
+        "source_track_results_sha256": "724d1fa367dcc57d07d4f6959a1cec40b98ddc6350179fc55248586fa1da6a3e",
+        "tasks": {
+          "boolq": {
+            "n": 3270,
+            "balanced_accuracy_percent": 59.1,
+            "classes": [
+              "no",
+              "yes"
+            ],
+            "predicted_class_percent": [
+              41.3,
+              58.7
+            ],
+            "gold_class_percent": [
+              37.8,
+              62.2
+            ]
+          },
+          "mmlu_pl": {
+            "n": 14042,
+            "balanced_accuracy_percent": 25.0,
+            "classes": [
+              "A",
+              "B",
+              "C",
+              "D"
+            ],
+            "predicted_class_percent": [
+              22.3,
+              38.6,
+              31.0,
+              8.1
+            ],
+            "gold_class_percent": [
+              22.9,
+              24.7,
+              25.5,
+              26.9
+            ]
+          },
+          "belebele_pol": {
+            "n": 900,
+            "balanced_accuracy_percent": 25.39,
+            "classes": [
+              "A",
+              "B",
+              "C",
+              "D"
+            ],
+            "predicted_class_percent": [
+              69.1,
+              5.7,
+              20.9,
+              4.3
+            ],
+            "gold_class_percent": [
+              22.9,
+              27.9,
+              27.3,
+              21.9
+            ]
+          }
+        }
+      },
+      "1.09-k35000": {
+        "source_diagnostic_sha256": "85cf33056bd30a85cf5411568ba6e2098eb31bbe9493d4156d800e98d6edcc1b",
+        "source_track_results_sha256": "01885b60efca4b0b1605502e5a1d9f02af7ed1718c4255c8d7c9a0288cac36c0",
+        "tasks": {
+          "boolq": {
+            "n": 3270,
+            "balanced_accuracy_percent": 60.29,
+            "classes": [
+              "no",
+              "yes"
+            ],
+            "predicted_class_percent": [
+              49.2,
+              50.8
+            ],
+            "gold_class_percent": [
+              37.8,
+              62.2
+            ]
+          },
+          "mmlu_pl": {
+            "n": 14042,
+            "balanced_accuracy_percent": 24.7,
+            "classes": [
+              "A",
+              "B",
+              "C",
+              "D"
+            ],
+            "predicted_class_percent": [
+              71.4,
+              15.9,
+              11.2,
+              1.5
+            ],
+            "gold_class_percent": [
+              22.9,
+              24.7,
+              25.5,
+              26.9
+            ]
+          },
+          "belebele_pol": {
+            "n": 900,
+            "balanced_accuracy_percent": 25.64,
+            "classes": [
+              "A",
+              "B",
+              "C",
+              "D"
+            ],
+            "predicted_class_percent": [
+              78.8,
+              4.1,
+              14.9,
+              2.2
+            ],
+            "gold_class_percent": [
+              22.9,
+              27.9,
+              27.3,
+              21.9
+            ]
+          }
+        }
+      }
+    }
+  }
 }

README.md

@@ -3,6 +3,7 @@ language:
 - pl
 - en
 library_name: pytorch
+license: apache-2.0
 pipeline_tag: text-generation
 tags:
 - polish
@@ -14,26 +15,37 @@ tags:
 
 # Slayer1B 1.09
 
-**Arkadiusz Słota, Anna Sieczka, Piotr Styła / Fabryka AI (SlayerLab)**
+**Arkadiusz Słota, Anna Sieczka, Piotr Styła, Aleksander Ogrodzki, Paweł Puzio, Marcin Kuśmierski / Fabryka AI (SlayerLab)**
 
-**Historical execution snapshot, 11 November 2026, 10:45 UTC:** continued pretraining from Slayer1B 1.08 was running on eight B200 spot GPUs after an H200 segment and checkpoint recoveries. Operators reported update 27,140 at that snapshot; this is dated progress telemetry, not a current availability claim or the endpoint of the run. The public tracking plan then showed 40,000 updates, while the trainer retained its 80,000-update scheduling argument. Neither value establishes completed exposure. The later planning decision below supersedes the requirement to complete terminal learning-rate decay. This is not a completed model or inference-ready weight release.
+## Intermediate base checkpoints
 
-| Verified execution evidence | Scope |
-|---|---|
-| CPT start | Approximately 04:50 UTC, 10 November 2026, on eight B200 GPUs |
-| Historical H200 recovery | Operators verified fresh numerical control and checkpoint restores during the earlier H200 segment |
-| Compute at the historical 10:45 UTC snapshot | Eight B200 spot GPUs; training had continued beyond update 27,000; later availability requires operator confirmation |
-| Checksum-verified weights-only copies on the workstation | Updates 5,000, 10,000, 15,000, 20,000 and 25,000; not public weight releases |
-| Latest verified full-state workstation copy | Update 25,000; includes optimizer state, retained separately from weights-only copies; not a claim about the newest recovery checkpoint on the cloud disk |
+**Continued pretraining is in progress; these are base models, not instruction-tuned models.** The checkpoints below were taken in the constant-learning-rate phase. The main-base plan does not select an in-run terminal decay; the configured later schedule boundary still requires operator review. Allocation pauses do not mean the campaign is complete.
+
+| Update | Positions in this continuation | Lineage lower bound, positions | Native weights file |
+|---|---:|---:|---|
+| 5,000 | 5,242,880,000 | 23,991,510,554 | `wagi_5000.pt` |
+| 10,000 | 10,485,760,000 | 29,234,390,554 | `wagi_10000.pt` |
+| 15,000 | 15,728,640,000 | 34,477,270,554 | `wagi_15000.pt` |
+| 20,000 | 20,971,520,000 | 39,720,150,554 | `wagi_20000.pt` |
+| 25,000 | 26,214,400,000 | 44,963,030,554 | `wagi_25000.pt` |
+| 30,000 | 31,457,280,000 | 50,205,910,554 | `wagi_30000.pt` |
+| 35,000 | 36,700,160,000 | 55,448,790,554 | `wagi_35000.pt` |
+
+Files are managed in the separate [checkpoint archive](https://huggingface.co/SlayerLab/Slayer1B-1.09-checkpoints). **Archive snapshot, 12 November 2026, 14:53 UTC:** the repository is public following the owner's decision at 14:42 UTC. Upload of the eight native checkpoint files started at 14:50 UTC and is in progress. A file becomes available only when it appears in the archive's file list; verify its checksum and size before use. This main repository contains documentation during training.
+
+The archive distinguishes seven weights-only `.pt` files from the full update-35,000 recovery checkpoint `ckpt_35000.pt`, which also contains optimizer and queue state. These are native training checkpoints, not a Transformers `AutoModel` or safetensors release. Check the archive's exact revision, file list and checksums before use; its publication status is separate from the evaluation table below.
 
-The [CPT Track history](https://track.fabryka.ai/run/71cf98ec-7c60-54a8-ac47-0ea7a23f573a) provides changing public telemetry; consult it and the operators for later execution status and hardware availability. The planning update is a decision, not evidence that its future steps have executed. This card is not a completed 1.09 model, final quality verdict or inference-ready weight release.
+Token counts: **positions in this continuation = updates × 1,048,576**, counting each logical update once, not repeated compute after recovery. **Lineage lower bound = 18,863,257,682 inherited positions + positions in this continuation − 114,627,128 identified-copy positions**. The lineage figure is a source- and file-disjoint lower bound, **not globally unique tokens**. Exact evaluated safetensors identities in [EVALUATION.json](EVALUATION.json) are distinct from the native `.pt` file digests in the archive.
+
+
+**Execution receipt reported by the operators, 12 November 2026, 15:44 UTC:** training resumed at approximately **15:41 UTC** from update **38,200** on an eight-GPU B200 spot instance. Weights-only milestones through update **35,000** and a full update-35,000 recovery checkpoint are retained on the trusted workstation. The [CPT Track history](https://track.fabryka.ai/run/71cf98ec-7c60-54a8-ac47-0ea7a23f573a) provides later telemetry; a published point and a saved recovery checkpoint have separate receipts.
 
 
 ## Objective and starting point
 
 **Funding update — 11 November 2026, 15:29 UTC:** the operator reported **USD 500 received for continued main-base training**. This is an additional funding receipt, not a measurement of the current account balance or a funded final update count. The earlier USD 425 campaign envelope belongs to the historical 12:07 UTC decision. The objective remains maximum useful affordable exposure with constant learning rates after warmup and no planned in-run terminal decay, retaining cost guards and the operational reserve for a recoverable checkpoint, transfer and safe shutdown. An extension approaching the configured decay boundary requires review. Stop at a verified full checkpoint when the real budget or allocation requires it; a 5,000-update boundary is not required. Positions per dollar measure operating efficiency, not model quality.
 
-**Operator-recipe history and confirmed adoption — 11 November 2026:** at **15:41 UTC**, the paired recipe files selecting `KROKI=120000` had been installed on the workstation while the running trainer still used 80,000. Operators subsequently confirmed adoption at **16:23:47 UTC**, resuming from **update 29,800 on H200** with actual runtime arguments **`--steps 120000 --ckpt-every 200`**, constant LR **0.0006** and configured WSD decay beginning at update **96,000**. This is a dated runtime-adoption receipt, not a claim of current hardware availability. Separately, Track's displayed **80,000-update plan** is a revisable budget estimate, not the trainer's scheduling argument or a promised funded endpoint. Neither 80,000 nor 120,000 updates, a completion date or over-100B globally unique exposure is guaranteed.
+**Recipe adoption — 11 November 2026:** operators confirmed the new recipe at **16:23:47 UTC**, resuming from **update 29,800 on H200**, with **`--steps 120000 --ckpt-every 200`**, LR **0.0006** and configured WSD decay beginning at update **96,000**. The operational target is **80,000 updates**, corresponding to **102,634,710,554 source- and file-disjoint lineage positions** under the stated accounting. The 120,000-update argument controls the scheduler. Allocation and funding can interrupt progress before the target.
 
 A separate, continuously curated option targets approximately **10B tokens of quality-selected “gold” data at context 4,096**. It remains isolated from main-base inputs and evaluation material, with its own manifest and licensing/admission gate; books selected with Arkadiusz Słota require title-level license approval. A later finishing run with decay needs its own run and budget decision. Preliminary exclusion lists are not a complete packed 10B corpus or a release, and quality-based selection is not evidence of a measured quality gain. See the [data card](DATA_CARD.md) for the dated inventory and isolation boundary.
 
@@ -43,7 +55,7 @@ The unchanged architecture is specified as **1,073,912,851 parameters**, 20 laye
 
 Architecture and tokenizer lineage includes [SlayerLab/Slayer1B-1.07](https://huggingface.co/SlayerLab/Slayer1B-1.07/blob/a090b54e658fb57dc5166c1fcc90cc6bf5c80106/README.md). The selected 1.08 base and the separate historical 1.08b ablation are different checkpoints; the continuation described here starts from 1.08.
 
-The [training configuration](TRAINING_CONFIGURATION.md) distinguishes the approved 120,000-update scheduling recipe, its confirmed runtime adoption, Track's separate budget estimate and the actual budget-limited final length. Completed 1.08 and the abandoned first attempt remain separate historical evidence.
+The [training configuration](TRAINING_CONFIGURATION.md) records the 120,000-update scheduling recipe and its adoption. The operational target of 80,000 updates and the actual completed count are separate fields. Completed 1.08 and the abandoned first attempt retain their own historical records.
 
 ## Current CPT configuration and budget-limited endpoint
 
@@ -52,12 +64,12 @@ The [training configuration](TRAINING_CONFIGURATION.md) distinguishes the approv
 | Selected scheduling argument | `KROKI=120000`; confirmed runtime `--steps 120000` adoption at 11 November, 16:23:47 UTC, resuming from update 29,800 on H200; configured decay begins at update 96,000 |
 | New-source allocation | **20% PL / 80% EN target**; not a measured cumulative model-language ratio |
 | Optimizer / schedule | Muon peak 0.02 plus AdamW peak 0.0006; 2,000-update warmup, then constant learning rates under the main-base plan; adoption receipt confirms LR 0.0006. The configured final-20% WSD decay to 0.00006 remains at update 96,000; an extension approaching it requires review |
-| Checkpoints | Runtime `--ckpt-every 200` confirmed at the 16:23:47 UTC adoption; weights-only checkpoints at each 5,000-update boundary; full-state and weights-only workstation copies verified through update 25,000 |
+| Checkpoints | Runtime `--ckpt-every 200` confirmed at the 16:23:47 UTC adoption; weights-only milestones every 5,000 updates. Native milestone files and full-state recovery files are distinguished in the separate checkpoint archive |
 | Code | Launcher v5 SHA256 `796c71016c1e167bfd3ca98b78518f91c3502a96eac233dd7186ef6639fcc656`; exact command and converter pins in the configuration document |
 
 Operators must review actual available funds, prices and allocation changes without assuming that the decay boundary is permanently unaffordable. An extended run must be reviewed before entering decay; do not shorten `KROKI` to a budget estimate and bring decay forward. Runtime changes need the operator's approved restart procedure and receipt. The trainer and v5 launcher are unchanged; the operator wrapper's argument ceiling was separately approved for the new recipe.
 
-The reviewed inherited count is **18,863,257,682 source/file-disjoint positions**, not a globally content-deduplicated history or total historical compute. The conservative c01 duplicate correction is **114,627,128 positions**, a sum of two exclusion-list counts rather than a measured union. Historical content-level near-duplicates remain unresolved. The separate requirement to prepare **over 100B positions of base-data inputs** is not achieved by adding inherited counts to new training steps; the [data card](DATA_CARD.md) reports what is actually prepared.
+The reviewed inherited count is **18,863,257,682 source/file-disjoint positions**. The conservative c01 correction is **114,627,128 positions**, computed from two exclusion-list counts. The training target is **over 100B source- and file-disjoint positions in the 1.00–1.09 lineage**; the formula reaches that threshold at update **77,488**, and the 80,000-update plan corresponds to approximately **102.6B**. Historical content-level near-duplicates remain unresolved. The [data card](DATA_CARD.md) separately records prepared input capacity.
 
 ## Initial CPT execution evidence
 
@@ -97,47 +109,76 @@ The last received public sample before the pause is update **450**, training los
 
 Fresh numerical control and step-zero `INIT-FROM` preflight were verified for the CPT start described above. The earlier permission for the abandoned slot was not reused as that evidence. Any later physical-host restart still requires its own numerical control; successful start does not prove a checkpoint restore.
 
-Working checkpoints are not published as Hugging Face weights. Operators independently verified weights-only workstation copies at updates 5,000, 10,000, 15,000, 20,000 and 25,000, and a separate full-state workstation copy at update 25,000. The later full 30,000 transfer was interrupted in `.tmp`; benchmark conversion is not evidence of a completed full-state workstation copy. These transfer records are distinct from newer recovery checkpoints on the node and from evaluated working updates through 30,000 below. Weights-only copies are retained, while the full workstation recovery copy is replaced by the next verified transfer. This retention policy does not eliminate exposure to a cloud-disk failure between transfers.
+Weights-only milestone exports and full-state recovery checkpoints serve different purposes. The linked archive distinguishes both kinds; this main repository hosts neither during training. A milestone file does not establish that the newest complete optimizer state was transferred or that a cloud allocation is still running.
 
 ## Data and exposure
 
-The first CPT reference, **c01**, contains **5,347,737,900 input positions** and TRAIN SHA256 `3341080241aeabd1e03c77530a6d735cb9501ecd788fba60ea065a944cd09ee1`. One 5,000-update stage processes **5,242,880,000 positions**, less than that input capacity. The separately approved c01–c06 queue supplies six stages through update 30,000; only the earlier stages have been consumed. Available input capacity, processed positions and historical content uniqueness remain distinct. See [the data card](DATA_CARD.md) for the queue identities, source changes and c01 duplicate correction.
+The first CPT reference, **c01**, contains **5,347,737,900 input positions** and TRAIN SHA256 `3341080241aeabd1e03c77530a6d735cb9501ecd788fba60ea065a944cd09ee1`. One 5,000-update stage processes **5,242,880,000 positions**, less than that input capacity. The historical c01–c06 queue reference supplied six stages through update 30,000; it is not the whole later prepared pool. Available input capacity, processed positions and historical content uniqueness remain distinct. See [the data card](DATA_CARD.md) for dated queue identities, source changes and the c01 duplicate correction.
 
 The PL/EN balance will follow the approved fresh-data inventory. The user explicitly allows a different ratio; source buckets still do not establish the language share of actually consumed windows. Missing ancestral data must be reported as a limitation, not silently counted as verified unique exposure.
 
 New material requires licensing, PII filtering, contamination scans, grouped held-outs and deduplication against the base model's full data ancestry. A 1% document holdout is not necessarily 1% of tokens. Historical exposure in the starting weights cannot be undone by filtering new inputs; retrospective benchmark hits must be handled explicitly in evaluation denominators.
 
-## Standard benchmark measurements — updates 5,000–30,000
+## Standard benchmark measurements — updates 5,000–35,000
+
+We thank **BenchTeam / Fabryka AI** for benchmark implementations, protocol work and evaluation proposals. Its independent [full benchmark report](https://github.com/slayerlabs/benchmarks-fabryka-ai/blob/8b547007526971bd2fc35dfe56ff2c7d9b802568/research/results/full-run-2026-11/README.md) is distinct from the **SlayerLab measurements below**, published in the [CPT Track history](https://track.fabryka.ai/run/71cf98ec-7c60-54a8-ac47-0ea7a23f573a). The seven listed 1.09 checkpoints are intermediate weights of the ongoing continuation, not its eventual budget-limited endpoint.
+
+The same nine full splits were scored zero-shot in float32 for 1.07 update 22,327, 1.08 update 1,709 and 1.09 updates 5,000, 10,000, 15,000, 20,000, 25,000, 30,000 and 35,000. Evaluator SHA, dataset revisions, file digests, task protocols and sample counts match across these rows. Scores are percentages with **pointwise 95% confidence intervals**; the named metric is accuracy (`acc`) or length-normalized accuracy (`acc_norm`). WinoGrande uses partial-sentence likelihood; MMLU-PL is micro-averaged. Polish HellaSwag and MMLU use the OpenGPTX task definitions.
 
-We thank **BenchTeam / Fabryka AI** for benchmark implementations, protocol work and evaluation proposals. Its independent [full benchmark report](https://github.com/slayerlabs/benchmarks-fabryka-ai/blob/8b547007526971bd2fc35dfe56ff2c7d9b802568/research/results/full-run-2026-11/README.md) is distinct from the **SlayerLab measurements below**, published in the [CPT Track history](https://track.fabryka.ai/run/71cf98ec-7c60-54a8-ac47-0ea7a23f573a). Updates 5,000, 10,000, 15,000, 20,000, 25,000 and 30,000 are distinct intermediate working checkpoints of the ongoing 1.09 run, not its eventual budget-limited endpoint or inference-ready weight releases.
+| Task | Metric | n | 1.07 | 1.08 | 1.09, update 5,000 | 1.09, update 10,000 | 1.09, update 15,000 | 1.09, update 20,000 | 1.09, update 25,000 | 1.09, update 30,000 | 1.09, update 35,000 |
+|---|---|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|
+| ARC-Challenge | acc_norm | 1,172 | 27.30 [24.74, 29.86] | 30.03 [27.39, 32.68] | 29.35 [26.79, 32.00] | 30.46 [27.82, 33.11] | 32.51 [29.86, 35.24] | 33.11 [30.46, 35.84] | 33.79 [31.06, 36.52] | 33.02 [30.29, 35.67] | 35.15 [32.42, 37.88] |
+| HellaSwag | acc_norm | 10,042 | 43.38 [42.43, 44.34] | 43.42 [42.43, 44.39] | 46.53 [45.55, 47.52] | 48.93 [47.97, 49.91] | 49.77 [48.80, 50.76] | 50.38 [49.39, 51.35] | 51.17 [50.18, 52.14] | 52.00 [51.02, 52.98] | 52.48 [51.50, 53.46] |
+| PIQA | acc_norm | 1,838 | 66.97 [64.80, 69.15] | 66.21 [64.15, 68.34] | 68.28 [66.21, 70.40] | 68.99 [66.92, 71.11] | 68.99 [66.87, 71.06] | 70.78 [68.72, 72.85] | 71.55 [69.48, 73.61] | 70.67 [68.61, 72.74] | 71.27 [69.15, 73.29] |
+| WinoGrande | acc | 1,267 | 53.59 [50.83, 56.35] | 53.67 [50.91, 56.43] | 54.93 [52.17, 57.70] | 54.93 [52.17, 57.62] | 56.75 [54.06, 59.43] | 55.01 [52.25, 57.77] | 55.88 [53.12, 58.64] | 57.14 [54.46, 59.83] | 56.75 [54.06, 59.43] |
+| BoolQ | acc | 3,270 | 58.29 [56.61, 59.97] | 59.97 [58.26, 61.59] | 61.80 [60.18, 63.49] | 61.13 [59.45, 62.78] | 59.88 [58.20, 61.53] | 51.41 [49.69, 53.12] | 52.20 [50.52, 53.91] | 60.67 [58.99, 62.35] | 59.88 [58.20, 61.56] |
+| OpenBookQA | acc_norm | 500 | 33.60 [29.40, 37.80] | 34.80 [30.60, 39.00] | 35.60 [31.40, 40.00] | 34.40 [30.20, 38.60] | 36.00 [31.80, 40.20] | 35.20 [31.00, 39.40] | 37.20 [32.80, 41.40] | 36.20 [32.00, 40.40] | 34.40 [30.40, 38.60] |
+| Belebele, Polish | acc | 900 | 26.00 [23.11, 28.89] | 24.00 [21.22, 26.89] | 24.11 [21.33, 26.89] | 24.11 [21.33, 27.00] | 22.56 [19.89, 25.33] | 24.67 [21.89, 27.56] | 25.33 [22.44, 28.22] | 24.44 [21.67, 27.33] | 24.33 [21.56, 27.11] |
+| HellaSwag-PL | acc_norm | 10,042 | 43.83 [42.84, 44.79] | 43.90 [42.90, 44.87] | 44.73 [43.76, 45.71] | 45.51 [44.51, 46.48] | 44.18 [43.21, 45.16] | 44.24 [43.26, 45.22] | 43.94 [42.94, 44.91] | 45.75 [44.77, 46.73] | 46.63 [45.68, 47.63] |
+| MMLU-PL | acc | 14,042 | 24.01 [23.30, 24.71] | 23.04 [22.34, 23.74] | 24.10 [23.39, 24.82] | 23.32 [22.62, 24.03] | 23.50 [22.80, 24.21] | 23.86 [23.17, 24.58] | 24.26 [23.57, 24.96] | 24.71 [23.99, 25.42] | 23.28 [22.58, 23.97] |
 
-The same nine full splits were scored zero-shot in float32 for 1.07 update 22,327, 1.08 update 1,709 and 1.09 updates 5,000, 10,000, 15,000, 20,000, 25,000 and 30,000. Evaluator SHA, dataset revisions, file digests, task protocols and sample counts match across these rows. Scores are percentages with **pointwise 95% confidence intervals**; the named metric is accuracy (`acc`) or length-normalized accuracy (`acc_norm`). WinoGrande uses partial-sentence likelihood; MMLU-PL is micro-averaged. Polish HellaSwag and MMLU use the OpenGPTX task definitions.
+Reference accuracies are context, not additional trained models or release thresholds:
 
-| Task | Metric | n | 1.07 | 1.08 | 1.09, update 5,000 | 1.09, update 10,000 | 1.09, update 15,000 | 1.09, update 20,000 | 1.09, update 25,000 | 1.09, update 30,000 |
-|---|---|---:|---:|---:|---:|---:|---:|---:|---:|---:|
-| ARC-Challenge | acc_norm | 1,172 | 27.30 [24.74, 29.86] | 30.03 [27.39, 32.68] | 29.35 [26.79, 32.00] | 30.46 [27.82, 33.11] | 32.51 [29.86, 35.24] | 33.11 [30.46, 35.84] | 33.79 [31.06, 36.52] | 33.02 [30.29, 35.67] |
-| HellaSwag | acc_norm | 10,042 | 43.38 [42.43, 44.34] | 43.42 [42.43, 44.39] | 46.53 [45.55, 47.52] | 48.93 [47.97, 49.91] | 49.77 [48.80, 50.76] | 50.38 [49.39, 51.35] | 51.17 [50.18, 52.14] | 52.00 [51.02, 52.98] |
-| PIQA | acc_norm | 1,838 | 66.97 [64.80, 69.15] | 66.21 [64.15, 68.34] | 68.28 [66.21, 70.40] | 68.99 [66.92, 71.11] | 68.99 [66.87, 71.06] | 70.78 [68.72, 72.85] | 71.55 [69.48, 73.61] | 70.67 [68.61, 72.74] |
-| WinoGrande | acc | 1,267 | 53.59 [50.83, 56.35] | 53.67 [50.91, 56.43] | 54.93 [52.17, 57.70] | 54.93 [52.17, 57.62] | 56.75 [54.06, 59.43] | 55.01 [52.25, 57.77] | 55.88 [53.12, 58.64] | 57.14 [54.46, 59.83] |
-| BoolQ | acc | 3,270 | 58.29 [56.61, 59.97] | 59.97 [58.26, 61.59] | 61.80 [60.18, 63.49] | 61.13 [59.45, 62.78] | 59.88 [58.20, 61.53] | 51.41 [49.69, 53.12] | 52.20 [50.52, 53.91] | 60.67 [58.99, 62.35] |
-| OpenBookQA | acc_norm | 500 | 33.60 [29.40, 37.80] | 34.80 [30.60, 39.00] | 35.60 [31.40, 40.00] | 34.40 [30.20, 38.60] | 36.00 [31.80, 40.20] | 35.20 [31.00, 39.40] | 37.20 [32.80, 41.40] | 36.20 [32.00, 40.40] |
-| Belebele, Polish | acc | 900 | 26.00 [23.11, 28.89] | 24.00 [21.22, 26.89] | 24.11 [21.33, 26.89] | 24.11 [21.33, 27.00] | 22.56 [19.89, 25.33] | 24.67 [21.89, 27.56] | 25.33 [22.44, 28.22] | 24.44 [21.67, 27.33] |
-| HellaSwag-PL | acc_norm | 10,042 | 43.83 [42.84, 44.79] | 43.90 [42.90, 44.87] | 44.73 [43.76, 45.71] | 45.51 [44.51, 46.48] | 44.18 [43.21, 45.16] | 44.24 [43.26, 45.22] | 43.94 [42.94, 44.91] | 45.75 [44.77, 46.73] |
-| MMLU-PL | acc | 14,042 | 24.01 [23.30, 24.71] | 23.04 [22.34, 23.74] | 24.10 [23.39, 24.82] | 23.32 [22.62, 24.03] | 23.50 [22.80, 24.21] | 23.86 [23.17, 24.58] | 24.26 [23.57, 24.96] | 24.71 [23.99, 25.42] |
+| Reference strategy | Scope | Accuracy (%) |
+|---|---|---:|
+| Uniform choice among four answers | Four-option reference for ARC-Challenge, HellaSwag/PL, OpenBookQA, Belebele-PL and MMLU-PL | 25.00 |
+| Uniform choice among two answers | BoolQ, PIQA and WinoGrande | 50.00 |
+| Always answer yes | BoolQ validation majority class: 2,033 of 3,270 labels | 62.17 |
+
+The majority-class reference is different from uniform guessing. A score near a reference does not prove that checkpoints are equal or that a model has no knowledge; individual confidence intervals are not paired tests. Class-balanced accuracy and answer-frequency diagnostics, when available, are auxiliary and do not replace the primary metrics above.
 
 Intervals use the percentile bootstrap with 10,000 replicates and seed 20261110. The updated approved aggregate does not supply the random-stream derivation or CI-generator SHA; the previous aggregate documented model-specific streams. **Items are treated as independent; subject, template and source clusters are not modeled.** These are intervals for individual results, not paired differences or training-seed variance; overlap is not a test of a difference.
 
 The [machine-readable aggregate record](EVALUATION.json) supplies exact values, intervals, counts, dataset revisions and file digests, checkpoint and evaluator identities, and metric selection, without prompts, answers or per-item predictions. ARC-Challenge, OpenBookQA and MMLU-PL use test splits; HellaSwag, PIQA, WinoGrande, BoolQ and HellaSwag-PL use validation splits; Belebele uses the full Polish evaluation set.
 
-The approved aggregate SHA-256 is `2d257a95523605b06aa964cb974f65d86ab025875cc1f87dc98b242b90729440`. All six earlier model records retain their scores, intervals, counts and dataset pins; the distinct update-25,000 and update-30,000 records extend that comparison without replacing the independent two-task protocol below.
+The approved aggregate SHA-256 is `d91456c7480f8031288afd5056c002ce68ef7d7751556d222ae54d7ab6fc648a`. All eight earlier model records retain their scores, intervals, counts and dataset pins; the update-35,000 record extends the nine-task comparison without replacing the independent two-task protocol below.
+
+**Descriptive results; lineage contamination checks are incomplete.** New-input chunks c03 onward were cleaned against these nine Track tasks, c01–c02 were only partially checked, and the inherited 1.00–1.08 data was not scanned. This is not a guarantee that these tasks are held out from the model's whole training history. Belebele and MMLU-PL are near their four-choice 25% random baseline; this is not proof of absent knowledge. No overall average, ranking or release threshold is inferred. Different evaluators must not be silently substituted.
 
-**Descriptive results; contamination scan of the training lineage against these tasks is pending.** Belebele and MMLU-PL are near their four-choice 25% random baseline. No overall average, ranking or release threshold is inferred. Numbers from different evaluators must not be silently substituted: prompting, BOS handling, normalization and software versions can change the result.
+MMLU-PL and Belebele-PL score the answer letter A–D given listed options; ARC-Challenge, HellaSwag/PL and PIQA score the answer text. This protocol difference is observed. Its contribution to the results requires a separate comparison.
 
 There is no replay-only control for this continuation, and initialization, exposure and schedule differ between checkpoints. These measurements do not isolate a causal effect of the new datasets or quantify training-seed variance. Earlier internal diagnostics are separate historical records, not a criterion for naming or publishing this model.
 
+### Auxiliary class-balanced diagnostic
+
+Balanced accuracy is the unweighted mean of recall across gold-answer classes. It was collected after the primary measurements. The values below are descriptive; their source does not provide confidence intervals. They retain a separate method and source digest in [EVALUATION.json](EVALUATION.json).
+
+| Update | BoolQ balanced accuracy (%) | MMLU-PL balanced accuracy (%) | Belebele-PL balanced accuracy (%) |
+|---|---:|---:|---:|
+| 5,000 | 53.39 | 25.34 | 25.60 |
+| 10,000 | 52.42 | 25.09 | 25.40 |
+| 15,000 | 53.98 | 24.99 | 24.42 |
+| 20,000 | 57.34 | 25.06 | 24.91 |
+| 25,000 | 57.44 | 24.77 | 26.25 |
+| 30,000 | 59.10 | 25.00 | 25.39 |
+| 35,000 | 60.29 | 24.70 | 25.64 |
+
+The corresponding uniform-choice reference is 50% for BoolQ and 25% for the two four-class tasks. These auxiliary values support inspection of class imbalance. They leave the primary metrics and their published intervals unchanged; a checkpoint-difference claim requires a separate paired analysis.
+
+
 ## Additional GSM8K and SST-2 measurements — update 10,000
 
-SlayerLab measured this two-task subset on 11 November 2026 using the [pinned BenchTeam code](https://github.com/slayerlabs/benchmarks-fabryka-ai/tree/8b547007526971bd2fc35dfe56ff2c7d9b802568), lm-eval 0.4.9.1, the `slayer` adapter, float32, batch 1 and context limit 4,096. These are full evaluated splits, not a full 50-task run or a copy of the community's independent report. **This separate two-task protocol measures 1.09 update 10,000 only; it does not supply GSM8K or SST-2 results for updates 15,000, 20,000, 25,000 or 30,000.** Update 10,000 is an internal working checkpoint, not a published or inference-ready 1.09 weight release.
+SlayerLab measured this two-task subset on 11 November 2026 using the [pinned BenchTeam code](https://github.com/slayerlabs/benchmarks-fabryka-ai/tree/8b547007526971bd2fc35dfe56ff2c7d9b802568), lm-eval 0.4.9.1, the `slayer` adapter, float32, batch 1 and context limit 4,096. These are full evaluated splits, not a full 50-task run or a copy of the community's independent report. **This separate two-task protocol measures 1.09 update 10,000 only; it does not supply GSM8K or SST-2 results for updates 15,000, 20,000, 25,000, 30,000 or 35,000.** Releasing another checkpoint does not supply a missing measurement.
 
 | Task | Metric / prompting | n | 1.07 | 1.08 | 1.09, update 10,000 |
 |---|---|---:|---:|---:|---:|
@@ -150,6 +191,6 @@ Scores are percentages with reported pointwise 95% bootstrap intervals (10,000 d
 
 ## Cost, licensing and limitations
 
-The total campaign ceiling remains **USD 425**, including GPU allocations, preparation, interruptions, storage and evaluation tracked by the resource meter. At the operator's 11 November, 10:39 UTC snapshot, the meter reported **approximately USD 239 spent**, leaving approximately **USD 186**; this is counter attribution, not a settled invoice or a permanently current balance. Hardware selection is based on effective cost per processed position and verified availability. Under the 12:07 planning decision, budget reviews preserve the existing cost guards and operational reserve, with no earmark for future gold-data training and no requirement to fund terminal decay in the main base. No higher spending limit or guaranteed completion date is implied.
+Continued training is budget-limited and may pause with spot availability. The funding receipt above is not the current balance or a guarantee of a final update count. Metered attribution is not a settled invoice; hardware choice must account for effective retained progress, interruptions, storage and evaluation, not only an advertised GPU-hour price. This page does not allocate a new budget to gold-data training or change the existing cost guards.
 
-The selected 1.08 base declares Apache-2.0; source-text rights and future publication still require separate approval. Privacy filters are heuristic; generated text may be inaccurate, biased, offensive or personal. No reasoning quality, instruction-following ability or professional reliability has been demonstrated for the revised 1.09.
+The model's declared license is **Apache-2.0**, aligned with the selected 1.08 base. Source-text rights and dataset publication require separate review; a model license does not grant rights to redistribute its training texts. Privacy filters are heuristic; generated text may be inaccurate, biased, offensive or personal. No reasoning quality, instruction-following ability or professional reliability has been demonstrated for the revised 1.09.

TRAINING_CONFIGURATION.md

@@ -36,13 +36,13 @@ bash /workspace/kod/uruchom_109.sh
 
 Set `START`, `GPU`, `S`, `KROKI` and `CKPT_EVERY` explicitly. The operator-approved recipe selects **`KROKI=120000`**, moving the 20% WSD decay boundary to **update 96,000**. The security reviewer approved the paired argument-file and operator-wrapper ceiling change on 11 November at 15:40 UTC; the operator installed them on the workstation at 15:41 UTC. Adoption was then independently verified after the **16:23:47 UTC restart from update 29,800**: runtime `--steps 120000`, `--ckpt-every 200` and learning rate **0.0006**, with the change from 80,000 to 120,000 recorded by the launcher. The trainer and v5 launcher code are unchanged. This confirms the approved schedule in that recovery, not current hardware availability or a funded final update count.
 
-The launcher permits this change only before `min(0.8 × old steps, 0.8 × new steps)`: for 80,000 → 120,000, before **update 64,000**. Review actual available funds, prices and allocation availability without assuming any decay boundary is permanently unaffordable. If more funding or lower prices make update 96,000 reachable, review the plan before it is reached to preserve the main-base no-decay objective. Do not shorten `KROKI` to a budget estimate and move decay earlier. Track's displayed **80,000-update plan** is a separate revisable estimate, not this scheduling argument or a guaranteed endpoint.
+The launcher permits this change only before `min(0.8 × old steps, 0.8 × new steps)`: for 80,000 → 120,000, before **update 64,000**. Operators review funds, prices and allocation availability before entering a later schedule phase. The operational target is **80,000 updates**, corresponding to **102,634,710,554 source- and file-disjoint lineage positions** under the accounting below. The scheduler uses 120,000 updates and starts its configured decay at update 96,000. Reducing the scheduling argument changes that boundary and requires a separate operator decision.
 
 Control values remain microbatch 8 × accumulation 4 × eight ranks, context 4,096, Muon peak 0.02 / AdamW peak 0.0006 and 2,000-update warmup. The configured final 20% decay to 0.00006 remains in the scheduler but is not the main-base objective. Full-state checkpoints used 400-update intervals historically; the verified 11 November, 14:53 UTC recovery used **200**, following the earlier cadence decision for short spot allocations. Each 5,000-update boundary also retains a weights-only checkpoint. Cadence is an operational choice recorded separately from the learning-rate schedule. Target source allocation is 20% PL / 80% EN, not the cumulative model-language ratio.
 
 `--init-from` imports weights/config from 1.08 into this new experiment with a fresh optimizer. `--resume` restores this experiment's own state; it must not resume the abandoned random-initialized attempt. A verified new run and a verified recovery are separate receipts.
 
-Each update processes **1,048,576 token positions**. Scheduling arguments, Track projections, prepared input capacity, reported progress and the latest saved checkpoint must not be substituted for one another. Recovery may repeat updates after the newest saved checkpoint; those repetitions are real compute but not new distinct data. The [data card](DATA_CARD.md) records the dated prepared-input inventory and the separate over-100B preparation requirement. Inherited source/file-disjoint counts are not a measurement of global content uniqueness.
+Each update processes **1,048,576 token positions**. The lineage lower bound is **18,863,257,682 + updates × 1,048,576 − 114,627,128 positions**; it first exceeds 100B at update **77,488**. It is source- and file-disjoint accounting, with historical content-level overlap unresolved. Operational targets, scheduler arguments, prepared input capacity, reported progress and saved recovery state have separate receipts. Recovery can repeat compute after a saved checkpoint. The [data card](DATA_CARD.md) records the prepared-input inventory and its separate preparation requirements.
 
 ## Separate future gold-data run — not funded or executed
 
@@ -59,7 +59,7 @@ Inherited-count footprint: counter SHA256 `abb39b8ce019c1a18a19ccf0ef4c32977c0c5
 
 Operators confirmed the 04:50 UTC start: fresh 1.07 hardware control within the 0.001 tolerance, pinned 1.08 weights/config converted to the training format, step-zero `INIT-FROM`, fresh optimizer, world size eight, c01/queue/CL validation identities and `PRE-FLIGHT OK`. The inherited 1.08 checkpoint step 1,709 is source metadata, **not the new continuation's initial training counter**.
 
-The distinct [CPT Track run](https://track.fabryka.ai/run/71cf98ec-7c60-54a8-ac47-0ea7a23f573a) preserves this experiment's real updates. Its 02:27 UTC registration is not the 04:50 UTC training start. The historical 10 November, 19:23 UTC snapshot verified full recovery state at update 10,000, SHA256 `01ccb2bed25c43b43356c8a813f9bf3f8b55c481bce85087e60862001f1e3dea`, and its weights-only counterpart `fcfd2d3044d504706519fb6e1bd8d6dea959b162dac62e0f5967bbd6807df643`. The model card separately records the later verified full-state and weights-only workstation copies through update 25,000. Neither receipt asserts the newest cloud checkpoint, a final model release or final quality acceptance.
+The distinct [CPT Track run](https://track.fabryka.ai/run/71cf98ec-7c60-54a8-ac47-0ea7a23f573a) preserves the experiment's published updates. Its 02:27 UTC registration and the reported 04:50 UTC training start are separate events. **Operator receipt, 12 November 2026, 15:44 UTC:** training resumed at approximately **15:41 UTC** from update **38,200** on an eight-GPU B200 spot instance. The trusted workstation retained weights-only milestones through **35,000** and a full update-35,000 recovery checkpoint. **Archive snapshot, 14:53 UTC:** the separate archive is public and upload of its eight native files is in progress. The [model card](README.md) and the archive's actual file list describe availability.
 
 
 ## Historical code identities