Changes to Get Experiment Group Metric Settings

GET /projects/{project_id}/experiment-groups/{group_id}/metric-settings

Every recorded change to this endpoint, newest first. Back to the operation

Changed in 4 of the 23 revisions of this API.176

Sep 2, 20264bc0315072461
Aug 13, 2026344acdc36dbb1
Aug 6, 2026933e8c8e63666
Jul 26, 202605003e7835a2105

What changed

Sep 2, 20264bc0315072461See the full diff
- ●
  added the new Detects whether model output concretely violates one or more applicable configured output-policy rules. enum value to the //// response property for the response status
  response-property-enum-value-added
This revision also has 4 changes that name no endpoint, such as unreferenced schemas being removed. See the revision's changelog
Aug 13, 2026344acdc36dbb1See the full diff
- ○
  removed the Measures which documents or chunks retrieved were used by the model to generate a response and how much of the text in the retrieved chunks was used by the model to compose its response. enum value from the //// response property for the response status
  response-property-enum-value-removed
This revision also has 44 changes that name no endpoint, such as unreferenced schemas being removed. See the revision's changelog
Aug 6, 2026933e8c8e63666See the full diff
- ●
  added the new Assesses whether a chatbot interaction that includes audio left the user feeling satisfied and positive or frustrated and dissatisfied based on tone engagement and overall experience. enum value to the //// response property for the response status
  response-property-enum-value-added
- ●
  added the new Assesses whether a chatbot interaction that includes images or documents left the user feeling satisfied and positive or frustrated and dissatisfied based on tone engagement and overall experience. enum value to the //// response property for the response status
  response-property-enum-value-added
- ●
  added the new Measures how sexist audio input content might be perceived ranging from 0 to 1. enum value to the //// response property for the response status
  response-property-enum-value-added
- ●
  added the new Measures how sexist image or document input content might be perceived ranging from 0 to 1. enum value to the //// response property for the response status
  response-property-enum-value-added
- ●
  added the new Measures how well the LLM follows system instructions provided in the prompt. enum value to the //// response property for the response status
  response-property-enum-value-added
- ●
  added the new Measures how well the LLM follows system instructions provided in the prompt. enum value to the //// response property for the response status
  response-property-enum-value-added
This revision also has 8 changes that name no endpoint, such as unreferenced schemas being removed. See the revision's changelog
Jul 26, 202605003e7835a2105See the full diff
- ●
  added the new Detects a significant shift in the user's primary conversational goal or workflow during a session that includes audio. enum value to the //// response property for the response status
  response-property-enum-value-added
- ●
  added the new Detects a significant shift in the user's primary conversational goal or workflow during a session that includes images or documents. enum value to the //// response property for the response status
  response-property-enum-value-added
- ●
  added the new Detects whether the user successfully accomplished all of their goals in a session that includes audio. enum value to the //// response property for the response status
  response-property-enum-value-added
- ●
  added the new Detects whether the user successfully accomplished all of their goals in a session that includes images or documents. enum value to the //// response property for the response status
  response-property-enum-value-added
- ●
  added the new Measures how well the workflow's response aligns with ground truth that includes audio. enum value to the //// response property for the response status
  response-property-enum-value-added
- ●
  added the new Measures how well the workflow's response aligns with ground truth that includes images or documents. enum value to the //// response property for the response status
  response-property-enum-value-added
- ●
  added the new Measures the coherence of the reasoning process by evaluating the consistency and logical flow of reasoning steps with audio context. enum value to the //// response property for the response status
  response-property-enum-value-added
- ●
  added the new Measures the coherence of the reasoning process by evaluating the consistency and logical flow of reasoning steps with image or document context. enum value to the //// response property for the response status
  response-property-enum-value-added
- ●
  added the new Measures the potential presence of factual errors or inconsistencies in the model's response including audio content. enum value to the //// response property for the response status
  response-property-enum-value-added
- ●
  added the new Measures the potential presence of factual errors or inconsistencies in the model's response including image or document content. enum value to the //// response property for the response status
  response-property-enum-value-added
- ○
  the endpoint scheme security ClassicAPIKeyHeader was added to the API
  api-security-added
- ○
  removed the A measure of the model's own confusion in its output. Higher scores indicate higher uncertainty. enum value from the //// response property for the response status
  response-property-enum-value-removed
- ○
  removed the BLEU is a case-sensitive measurement of the difference between an model generation and target generation at the sentence-level. enum value from the //// response property for the response status
  response-property-enum-value-removed
- ○
  removed the Measures the perplexity of the prompt. Lower perplexity score is generally considered to be better because it means the model is less surprised by the text and can predict the next word in a sentence with higher accuracy. enum value from the //// response property for the response status
  response-property-enum-value-removed
- ○
  removed the ROUGE measures the unigram overlap between model generation and target generation as a single F-1 score. enum value from the //// response property for the response status
  response-property-enum-value-removed
This revision also has 15 changes that name no endpoint, such as unreferenced schemas being removed. See the revision's changelog

Of the 23 revisions, 2 have no diff computed.