| { |
| "v2": { |
| "vocab_size": 32000, |
| "pollution_summary": { |
| "handler_suffix": 57, |
| "and_camel_class": 43, |
| "long_camel": 3, |
| "uppercase_blob": 462, |
| "engineering_suffix": 11, |
| "very_long_token": 28 |
| }, |
| "unique_polluted_count": 551, |
| "polluted_examples": [ |
| [ |
| 3114, |
| "Handler" |
| ], |
| [ |
| 24052, |
| "ugmentationHumaineHandler" |
| ], |
| [ |
| 23384, |
| " DevOpsAndCloudInfrastructureHandler" |
| ], |
| [ |
| 24695, |
| "AndAnalyticsHandler" |
| ], |
| [ |
| 23259, |
| "dustrialEngineeringHandler" |
| ], |
| [ |
| 23384, |
| " DevOpsAndCloudInfrastructureHandler" |
| ], |
| [ |
| 24420, |
| "énarisationAndSt" |
| ], |
| [ |
| 24054, |
| " ImmobilierAndInvestissementHandler" |
| ], |
| [ |
| 24926, |
| " NeurosciencesAndSciences" |
| ], |
| [ |
| 22133, |
| " RésolutiondeConflitsAndMé" |
| ], |
| [ |
| 23385, |
| "LevelSystemsProgrammingHandler" |
| ], |
| [ |
| 24934, |
| "PrivacyAndComplianceHandler" |
| ], |
| [ |
| 23860, |
| "AndCorporateLawHandler" |
| ], |
| [ |
| 8046, |
| "GRES" |
| ], |
| [ |
| 30173, |
| "BNBUSDT" |
| ], |
| [ |
| 12631, |
| "ULTA" |
| ], |
| [ |
| 18722, |
| "IPORTS" |
| ], |
| [ |
| 4626, |
| "CRITIQUE" |
| ], |
| [ |
| 23259, |
| "dustrialEngineeringHandler" |
| ], |
| [ |
| 8914, |
| " Engineering" |
| ], |
| [ |
| 25336, |
| " DataEngineeringAndSQLAvancéHandler" |
| ], |
| [ |
| 24424, |
| "ScénarisationAndStorytellingHandler" |
| ], |
| [ |
| 23252, |
| " ManufacturingAndIn" |
| ], |
| [ |
| 23384, |
| " DevOpsAndCloudInfrastructureHandler" |
| ], |
| [ |
| 23385, |
| "LevelSystemsProgrammingHandler" |
| ], |
| [ |
| 30904, |
| " " |
| ], |
| [ |
| 24054, |
| " ImmobilierAndInvestissementHandler" |
| ], |
| [ |
| 23678, |
| " JournalismeAndInvestigationHandler" |
| ] |
| ], |
| "fertility": [ |
| { |
| "corpus": "EN", |
| "chars": 545, |
| "tokens": 150, |
| "tokens_per_100_chars": 27.5 |
| }, |
| { |
| "corpus": "FR", |
| "chars": 761, |
| "tokens": 184, |
| "tokens_per_100_chars": 24.2 |
| }, |
| { |
| "corpus": "code", |
| "chars": 1332, |
| "tokens": 585, |
| "tokens_per_100_chars": 43.9 |
| }, |
| { |
| "corpus": "tool_use_chatml", |
| "chars": 477, |
| "tokens": 182, |
| "tokens_per_100_chars": 38.2 |
| } |
| ], |
| "fr_top_test": { |
| "1_token": 140, |
| "2_token": 23, |
| "3_token": 20, |
| "4plus_token": 7, |
| "examples_high": [ |
| [ |
| "française", |
| 4 |
| ], |
| [ |
| "chômage", |
| 4 |
| ], |
| [ |
| "aujourd'hui", |
| 4 |
| ], |
| [ |
| "qu'est-ce", |
| 5 |
| ], |
| [ |
| "j'aime", |
| 4 |
| ], |
| [ |
| "n'est-ce", |
| 5 |
| ], |
| [ |
| "rejoindre", |
| 4 |
| ] |
| ], |
| "total": 190, |
| "pct_1tok": 73.7, |
| "pct_efficient": 85.8 |
| }, |
| "special_tokens": [ |
| [ |
| 0, |
| "<pad>", |
| true |
| ], |
| [ |
| 1, |
| "<bos>", |
| true |
| ], |
| [ |
| 2, |
| "<eos>", |
| true |
| ], |
| [ |
| 3, |
| "<unk>", |
| true |
| ], |
| [ |
| 4, |
| "<sep>", |
| true |
| ], |
| [ |
| 5, |
| "<|teach|>", |
| true |
| ], |
| [ |
| 6, |
| "<|evolve|>", |
| true |
| ], |
| [ |
| 7, |
| "<|query|>", |
| true |
| ], |
| [ |
| 8, |
| "<|code|>", |
| true |
| ], |
| [ |
| 9, |
| "<|/code|>", |
| true |
| ], |
| [ |
| 10, |
| "<|rust|>", |
| true |
| ], |
| [ |
| 11, |
| "<|python|>", |
| true |
| ], |
| [ |
| 12, |
| "<|system|>", |
| true |
| ], |
| [ |
| 13, |
| "<|user|>", |
| true |
| ], |
| [ |
| 14, |
| "<|archon|>", |
| true |
| ], |
| [ |
| 15, |
| "<|think|>", |
| true |
| ] |
| ] |
| }, |
| "v21": { |
| "vocab_size": 32006, |
| "pollution_summary": { |
| "handler_suffix": 57, |
| "and_camel_class": 43, |
| "long_camel": 3, |
| "uppercase_blob": 462, |
| "engineering_suffix": 11, |
| "very_long_token": 28 |
| }, |
| "unique_polluted_count": 551, |
| "polluted_examples": [ |
| [ |
| 23176, |
| "artupHandler" |
| ], |
| [ |
| 24917, |
| "tivesHandler" |
| ], |
| [ |
| 24695, |
| "AndAnalyticsHandler" |
| ], |
| [ |
| 25501, |
| " GénétiqueAndBiotechnologieHandler" |
| ], |
| [ |
| 23850, |
| "LawHandler" |
| ], |
| [ |
| 22929, |
| " ArchitectureAndUrbanisme" |
| ], |
| [ |
| 23675, |
| " JournalismeAndInvestigation" |
| ], |
| [ |
| 24421, |
| "ScénarisationAndSt" |
| ], |
| [ |
| 22133, |
| " RésolutiondeConflitsAndMé" |
| ], |
| [ |
| 25501, |
| " GénétiqueAndBiotechnologieHandler" |
| ], |
| [ |
| 23860, |
| "AndCorporateLawHandler" |
| ], |
| [ |
| 23385, |
| "LevelSystemsProgrammingHandler" |
| ], |
| [ |
| 24934, |
| "PrivacyAndComplianceHandler" |
| ], |
| [ |
| 3315, |
| "ORDRE" |
| ], |
| [ |
| 15231, |
| "VUSDT" |
| ], |
| [ |
| 11028, |
| "AUDE" |
| ], |
| [ |
| 23399, |
| "YOLO" |
| ], |
| [ |
| 24086, |
| "MPER" |
| ], |
| [ |
| 23251, |
| "dustrialEngineering" |
| ], |
| [ |
| 11602, |
| " Manufacturing" |
| ], |
| [ |
| 23252, |
| " ManufacturingAndIn" |
| ], |
| [ |
| 15546, |
| "Engineering" |
| ], |
| [ |
| 8914, |
| " Engineering" |
| ], |
| [ |
| 8212, |
| "|-------------|----------|------------" |
| ], |
| [ |
| 25501, |
| " GénétiqueAndBiotechnologieHandler" |
| ], |
| [ |
| 24167, |
| "ÃīnigmesAndRaisonnementLatéralHandler" |
| ], |
| [ |
| 8229, |
| "|-------------|----------|---------------|----------|" |
| ], |
| [ |
| 30904, |
| " " |
| ] |
| ], |
| "fertility": [ |
| { |
| "corpus": "EN", |
| "chars": 545, |
| "tokens": 150, |
| "tokens_per_100_chars": 27.5 |
| }, |
| { |
| "corpus": "FR", |
| "chars": 761, |
| "tokens": 184, |
| "tokens_per_100_chars": 24.2 |
| }, |
| { |
| "corpus": "code", |
| "chars": 1332, |
| "tokens": 585, |
| "tokens_per_100_chars": 43.9 |
| }, |
| { |
| "corpus": "tool_use_chatml", |
| "chars": 477, |
| "tokens": 141, |
| "tokens_per_100_chars": 29.6 |
| } |
| ], |
| "fr_top_test": { |
| "1_token": 140, |
| "2_token": 23, |
| "3_token": 20, |
| "4plus_token": 7, |
| "examples_high": [ |
| [ |
| "française", |
| 4 |
| ], |
| [ |
| "chômage", |
| 4 |
| ], |
| [ |
| "aujourd'hui", |
| 4 |
| ], |
| [ |
| "qu'est-ce", |
| 5 |
| ], |
| [ |
| "j'aime", |
| 4 |
| ], |
| [ |
| "n'est-ce", |
| 5 |
| ], |
| [ |
| "rejoindre", |
| 4 |
| ] |
| ], |
| "total": 190, |
| "pct_1tok": 73.7, |
| "pct_efficient": 85.8 |
| }, |
| "special_tokens": [ |
| [ |
| 0, |
| "<pad>", |
| true |
| ], |
| [ |
| 1, |
| "<bos>", |
| true |
| ], |
| [ |
| 2, |
| "<eos>", |
| true |
| ], |
| [ |
| 3, |
| "<unk>", |
| true |
| ], |
| [ |
| 4, |
| "<sep>", |
| true |
| ], |
| [ |
| 5, |
| "<|teach|>", |
| true |
| ], |
| [ |
| 6, |
| "<|evolve|>", |
| true |
| ], |
| [ |
| 7, |
| "<|query|>", |
| true |
| ], |
| [ |
| 8, |
| "<|code|>", |
| true |
| ], |
| [ |
| 9, |
| "<|/code|>", |
| true |
| ], |
| [ |
| 10, |
| "<|rust|>", |
| true |
| ], |
| [ |
| 11, |
| "<|python|>", |
| true |
| ], |
| [ |
| 12, |
| "<|system|>", |
| true |
| ], |
| [ |
| 13, |
| "<|user|>", |
| true |
| ], |
| [ |
| 14, |
| "<|archon|>", |
| true |
| ], |
| [ |
| 15, |
| "<|think|>", |
| true |
| ], |
| [ |
| 32000, |
| "<|im_start|>", |
| true |
| ], |
| [ |
| 32001, |
| "<|im_end|>", |
| true |
| ], |
| [ |
| 32002, |
| "<|assistant|>", |
| true |
| ], |
| [ |
| 32003, |
| "<|tool_call|>", |
| true |
| ], |
| [ |
| 32004, |
| "<|tool_result|>", |
| true |
| ], |
| [ |
| 32005, |
| "<|task_type|>", |
| true |
| ] |
| ] |
| }, |
| "v21_diff": { |
| "added_in_v21": [ |
| "<|assistant|>", |
| "<|im_end|>", |
| "<|im_start|>", |
| "<|task_type|>", |
| "<|tool_call|>", |
| "<|tool_result|>" |
| ], |
| "removed_in_v21": [] |
| } |
| } |