> ## Documentation Index > Fetch the complete documentation index at: https://docs.galtea.ai/llms.txt > Use this file to discover all available pages before exploring further. # Remove a human annotation > Clear the human score, reason and evaluator from an evaluation a judge scored, leaving it unclaimed. The AI score, reason and status are untouched. Where the judge produced a score, every reported number goes back to it. Where the judge produced none, as on a FAILED, SKIPPED or CANCELLED evaluation, the evaluation stops being reported altogether, so analytics, the product score, coverage and monitor results all drop it. The removal is permanent. The human score and reason are deleted, not archived, so getting them back means annotating the evaluation again. Refused on HUMAN_EVALUATION metrics, whose human score is the only score by design, on evaluations waiting for a human score, and on outdated evaluations, which accept no annotation at all. See [Evaluations](https://docs.galtea.ai/concepts/product/version/session/evaluation). ## OpenAPI ````yaml https://api.galtea.ai/openapi.json delete /evaluations/{id}/human-annotation openapi: 3.0.0 info: version: 1.0.0 title: Product Management Service API description: API documentation for Product Management Service contact: name: Galtea AI servers: - url: https://api.galtea.ai security: - bearerAuth: [] tags: [] externalDocs: description: Galtea Platform Documentation url: https://docs.galtea.ai paths: /evaluations/{id}/human-annotation: delete: tags: - evaluations summary: Remove a human annotation description: >- Clear the human score, reason and evaluator from an evaluation a judge scored, leaving it unclaimed. The AI score, reason and status are untouched. Where the judge produced a score, every reported number goes back to it. Where the judge produced none, as on a FAILED, SKIPPED or CANCELLED evaluation, the evaluation stops being reported altogether, so analytics, the product score, coverage and monitor results all drop it. The removal is permanent. The human score and reason are deleted, not archived, so getting them back means annotating the evaluation again. Refused on HUMAN_EVALUATION metrics, whose human score is the only score by design, on evaluations waiting for a human score, and on outdated evaluations, which accept no annotation at all. See [Evaluations](https://docs.galtea.ai/concepts/product/version/session/evaluation). operationId: removeHumanAnnotation parameters: - schema: type: string description: Evaluation ID example: eval_123 required: true description: Evaluation ID name: id in: path - $ref: '#/components/parameters/Prefer' responses: '200': description: >- Human annotation removed successfully. The body is the resource as its GET answers it. headers: Content-Location: schema: type: string description: >- The path of the resource in the body; absent when there is no body required: false description: >- The path of the resource in the body; absent when there is no body Preference-Applied: schema: type: string enum: - return=minimal description: 'Present when the API honored `Prefer: return=minimal`' required: false description: 'Present when the API honored `Prefer: return=minimal`' content: application/json: schema: $ref: '#/components/schemas/Evaluation' '204': description: >- No body: the client sent `Prefer: return=minimal`, or the read after the write failed. headers: Preference-Applied: schema: type: string enum: - return=minimal description: 'Present when the API honored `Prefer: return=minimal`' required: false description: 'Present when the API honored `Prefer: return=minimal`' '400': description: >- Evaluation has no human annotation, its metric is scored by a person, it is waiting for a human score, or it is outdated content: application/json: schema: $ref: '#/components/schemas/Error' '401': description: Unauthorized content: application/json: schema: $ref: '#/components/schemas/Error' '403': description: User does not have permission to annotate this evaluation content: application/json: schema: $ref: '#/components/schemas/Error' '404': description: Evaluation not found content: application/json: schema: $ref: '#/components/schemas/Error' security: - bearerAuth: [] components: parameters: Prefer: schema: $ref: '#/components/schemas/Prefer' required: false description: >- Send `return=minimal` to get no body back from a write and skip reading the resource: a create answers `201` with only `Location`, and an update answers `204`. name: Prefer in: header schemas: Evaluation: type: object properties: id: type: string example: eval_123 metricId: type: string example: metric_123 sessionId: type: string example: session_123 productId: type: string nullable: true example: product_123 description: Product the evaluated session belongs to userId: type: string nullable: true example: user_123 status: type: string enum: - PENDING - PENDING_HUMAN - SUCCESS - FAILED - SKIPPED - CANCELLED - OUTDATED example: SUCCESS inferenceResultId: type: string nullable: true example: ir_123 score: type: number nullable: true example: 0.95 reason: type: string nullable: true example: High quality response error: type: string nullable: true canRetry: type: boolean example: false creditsUsed: type: integer nullable: true example: 1 judgeInputTokens: type: integer nullable: true example: 4100 description: Tokens the judge received, prompt plus session content judgeOutputTokens: type: integer nullable: true example: 16800 description: Tokens the judge produced, reasoning included judgeReasoningTokens: type: integer nullable: true example: 12400 description: Reasoning tokens, a subset of judgeOutputTokens conversationSimulatorVersion: type: string nullable: true example: 1.0.0 humanEvaluatorId: type: string nullable: true description: User ID of the human evaluator humanEvaluatorStartedAt: type: string format: date-time nullable: true humanScore: type: number nullable: true description: Human-provided annotation score humanReason: type: string nullable: true description: Human-provided annotation reason humanEvaluatorFinishedAt: type: string nullable: true format: date-time description: Timestamp when human evaluation was submitted failedTurns: type: array items: type: string description: Conversation turns that failed monitorId: type: string nullable: true description: ID of the monitor that dispatched this evaluation monitorSettledAtTurn: type: integer nullable: true description: Session turn count when the monitor scored it monitorSamplingPercentage: type: number nullable: true description: Sampling percentage in effect when the monitor scored this session runId: type: string nullable: true description: ID of the run that launched this evaluation createdAt: type: string format: date-time deletedAt: type: string format: date-time nullable: true evaluatedAt: type: string format: date-time nullable: true metricLegacyAt: type: string format: date-time nullable: true passed: type: boolean nullable: true description: >- Whether the reported score (the reviewer's when set, else the judge's) reaches the metric's pass threshold. Null when the evaluation has no reported score, is outdated or deleted, is a long classifier input, or its metric is not loaded. metricPassThreshold: type: number minimum: 0 maximum: 1 description: >- The pass threshold that decides `passed`: the metric's, copied onto the evaluation when it was created. metricDisabledAt: type: string format: date-time nullable: true metricExcludedFromAnalyticsAt: type: string format: date-time nullable: true metricExcludedByUserId: type: string nullable: true testCaseLegacyAt: type: string format: date-time nullable: true scoreBreakdown: type: object nullable: true properties: version: type: integer model: type: string type: type: string enum: - yes_no - scale value: type: number confidence: type: number nullable: true pYes: type: number minimum: 0 maximum: 1 goodAnswer: type: string enum: - 'yes' - 'no' probabilities: type: object additionalProperties: type: number description: Per level index, the "does not apply" index included. levelScores: type: array items: type: number doesNotApply: type: boolean description: Whether the question had a "does not apply" level. required: - version - model - type - value - confidence description: >- CLASSIFIER only: what the classifier answered; null for every other source. confidence: type: number nullable: true minimum: 0 maximum: 1 example: 0.9 description: >- CLASSIFIER only: how sure the classifier was, from 0 to 1; null for every other source. Values do not compare across question types. longInput: type: boolean description: >- True for a CLASSIFIER evaluation whose input ran past what the classifier scores reliably (CLASSIFIER_RELIABLE_INPUT_TOKENS judge input tokens) and that no reviewer has scored; false for every other evaluation. required: - id - metricId - sessionId - productId - userId - status - inferenceResultId - score - reason - error - canRetry - creditsUsed - judgeInputTokens - judgeOutputTokens - judgeReasoningTokens - conversationSimulatorVersion - humanEvaluatorId - humanEvaluatorStartedAt - humanScore - humanReason - humanEvaluatorFinishedAt - failedTurns - monitorId - monitorSettledAtTurn - monitorSamplingPercentage - runId - deletedAt - evaluatedAt - metricLegacyAt - passed - metricPassThreshold - metricDisabledAt - metricExcludedFromAnalyticsAt - metricExcludedByUserId - testCaseLegacyAt - scoreBreakdown - confidence - longInput Error: type: object properties: error: type: string example: Error type message: type: string example: Error message description Prefer: type: string description: >- Send `return=minimal` to get no body back from a write and skip reading the resource: a create answers `201` with only `Location`, and an update answers `204`. example: return=minimal securitySchemes: bearerAuth: type: http scheme: bearer description: >- API key authorization. Pass your API key in the Authorization header as a Bearer token. Both new (`gsk_*`) and legacy (`gsk-`) API keys are accepted, e.g. `Authorization: Bearer gsk_...` or `Authorization: Bearer gsk-...`. ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.