This commit is contained in:
Sai Soundararaj 2025-07-01 17:00:35 -07:00
parent a9d8fdef90
commit 78ef9c605f
5 changed files with 125 additions and 25 deletions

View file

@ -9301,7 +9301,8 @@
"categorical_count",
"accuracy"
],
"title": "AggregationFunctionType"
"title": "AggregationFunctionType",
"description": "Types of aggregation functions for scoring results."
},
"BasicScoringFnParams": {
"type": "object",
@ -9309,13 +9310,15 @@
"type": {
"$ref": "#/components/schemas/ScoringFnParamsType",
"const": "basic",
"default": "basic"
"default": "basic",
"description": "The type of scoring function parameters, always basic"
},
"aggregation_functions": {
"type": "array",
"items": {
"$ref": "#/components/schemas/AggregationFunctionType"
}
},
"description": "Aggregation functions to apply to the scores of each row"
}
},
"additionalProperties": false,
@ -9323,7 +9326,8 @@
"type",
"aggregation_functions"
],
"title": "BasicScoringFnParams"
"title": "BasicScoringFnParams",
"description": "Parameters for basic scoring function configuration."
},
"BenchmarkConfig": {
"type": "object",
@ -9375,25 +9379,30 @@
"type": {
"$ref": "#/components/schemas/ScoringFnParamsType",
"const": "llm_as_judge",
"default": "llm_as_judge"
"default": "llm_as_judge",
"description": "The type of scoring function parameters, always llm_as_judge"
},
"judge_model": {
"type": "string"
"type": "string",
"description": "Identifier of the LLM model to use as a judge for scoring"
},
"prompt_template": {
"type": "string"
"type": "string",
"description": "(Optional) Custom prompt template for the judge model"
},
"judge_score_regexes": {
"type": "array",
"items": {
"type": "string"
}
},
"description": "Regexes to extract the answer from generated response"
},
"aggregation_functions": {
"type": "array",
"items": {
"$ref": "#/components/schemas/AggregationFunctionType"
}
},
"description": "Aggregation functions to apply to the scores of each row"
}
},
"additionalProperties": false,
@ -9403,7 +9412,8 @@
"judge_score_regexes",
"aggregation_functions"
],
"title": "LLMAsJudgeScoringFnParams"
"title": "LLMAsJudgeScoringFnParams",
"description": "Parameters for LLM-as-judge scoring function configuration."
},
"ModelCandidate": {
"type": "object",
@ -9441,19 +9451,22 @@
"type": {
"$ref": "#/components/schemas/ScoringFnParamsType",
"const": "regex_parser",
"default": "regex_parser"
"default": "regex_parser",
"description": "The type of scoring function parameters, always regex_parser"
},
"parsing_regexes": {
"type": "array",
"items": {
"type": "string"
}
},
"description": "Regex to extract the answer from generated response"
},
"aggregation_functions": {
"type": "array",
"items": {
"$ref": "#/components/schemas/AggregationFunctionType"
}
},
"description": "Aggregation functions to apply to the scores of each row"
}
},
"additionalProperties": false,
@ -9462,7 +9475,8 @@
"parsing_regexes",
"aggregation_functions"
],
"title": "RegexParserScoringFnParams"
"title": "RegexParserScoringFnParams",
"description": "Parameters for regex parser scoring function configuration."
},
"ScoringFnParams": {
"oneOf": [
@ -9492,7 +9506,8 @@
"regex_parser",
"basic"
],
"title": "ScoringFnParamsType"
"title": "ScoringFnParamsType",
"description": "Types of scoring function parameter configurations."
},
"EvaluateRowsRequest": {
"type": "object",
@ -10765,9 +10780,9 @@
"tool",
"tool_group"
],
"title": "ResourceType",
"const": "scoring_function",
"default": "scoring_function"
"default": "scoring_function",
"description": "The resource type, always scoring_function"
},
"description": {
"type": "string"
@ -10812,7 +10827,8 @@
"metadata",
"return_type"
],
"title": "ScoringFn"
"title": "ScoringFn",
"description": "A scoring function resource for evaluating model outputs."
},
"StringType": {
"type": "object",
@ -16105,20 +16121,23 @@
"type": "object",
"properties": {
"dataset_id": {
"type": "string"
"type": "string",
"description": "(Optional) The identifier of the dataset that was scored"
},
"results": {
"type": "object",
"additionalProperties": {
"$ref": "#/components/schemas/ScoringResult"
}
},
"description": "A map of scoring function name to ScoringResult"
}
},
"additionalProperties": false,
"required": [
"results"
],
"title": "ScoreBatchResponse"
"title": "ScoreBatchResponse",
"description": "Response from batch scoring operations on datasets."
},
"AlgorithmConfig": {
"oneOf": [