Skip to main content

​
Returns

Returns a list of Evaluation objects.

​
Example

​
Parameters

​
str | list[str]
default:"None"
The ID or IDs of the evaluation(s) to retrieve.
​
str | list[str]
default:"None"
The ID or IDs of the session(s) for which you want to list evaluations.
​
str
default:"None"
The ID of the version for which you want to list evaluations.
​
str | list[str]
default:"None"
The ID or IDs of the test case(s) for which you want to list evaluations.
​
str | list[str]
default:"None"
The ID or IDs of the dataset(s) for which you want to list evaluations.
Deprecated alias: test_id.
​
str | list[str]
default:"None"
The ID or IDs of the metric(s) for which you want to list evaluations.
​
str | list[str]
default:"None"
Filter by evaluation status(es). Valid values are 'PENDING', 'PENDING_HUMAN', 'SUCCESS', 'FAILED', 'SKIPPED', 'CANCELLED', 'OUTDATED'. Other string values are sent to the API unchanged, so values added in newer API versions can be used; the API rejects invalid ones.
​
str | int
default:"None"
Filter evaluations created at or after this timestamp. Accepts ISO 8601 string (e.g., '2024-01-01T00:00:00Z') or Unix timestamp in seconds. Can be used independently or together with to_created_at.
​
str | int
default:"None"
Filter evaluations created at or before this timestamp. Accepts ISO 8601 string (e.g., '2024-12-31T23:59:59Z') or Unix timestamp in seconds. Can be used independently or together with from_created_at.
​
string
Sort evaluations by creation date. Valid values are “asc” and “desc”.
​
string
Sort evaluations by score. Valid values are “asc” and “desc”.
​
int
default:"0"
Number of evaluations to skip before starting to collect results. Use for pagination.Default: 0 (start from the beginning)
​
int
default:"10000"
Maximum number of evaluations to return in a single request.Default: 10000
​
str | list[str]
default:"None"
Filter by the ID of the user who launched the evaluation. Null for a monitor-dispatched evaluation. This is an additional filter; it does not satisfy the scoping requirement below.
​
str | list[str]
default:"None"
The ID or IDs of the Runs that launched the evaluations to retrieve. A monitor-dispatched evaluation belongs to no run, so it never matches.
​
bool
default:"False"
Keyword-only. Include evaluations whose metric revision is excluded from analytics. These results keep their score and reason, but they do not feed any analytics number. See Metrics.
​
float
default:"None"
Keyword-only. Keep evaluations whose judge score is greater than this value, on the 0 to 1 scale. This is the judge’s score, not the human review. An evaluation with no judge score never matches. A value outside 0 to 1 raises ValueError.
​
float
default:"None"
Keyword-only. Keep evaluations whose judge score is greater than or equal to this value, on the 0 to 1 scale. This is the judge’s score, not the human review. An evaluation with no judge score never matches. A value outside 0 to 1 raises ValueError.
​
float
default:"None"
Keyword-only. Keep evaluations whose judge score is less than this value, on the 0 to 1 scale. This is the judge’s score, not the human review. An evaluation with no judge score never matches. A value outside 0 to 1 raises ValueError.
​
float
default:"None"
Keyword-only. Keep evaluations whose judge score is less than or equal to this value, on the 0 to 1 scale. This is the judge’s score, not the human review. An evaluation with no judge score never matches. A value outside 0 to 1 raises ValueError.
​
float
default:"None"
Keyword-only. Keep evaluations whose human review score is greater than this value, on the 0 to 1 scale. An evaluation with no human score never matches. A value outside 0 to 1 raises ValueError.
​
float
default:"None"
Keyword-only. Keep evaluations whose human review score is greater than or equal to this value, on the 0 to 1 scale. An evaluation with no human score never matches. A value outside 0 to 1 raises ValueError.
​
float
default:"None"
Keyword-only. Keep evaluations whose human review score is less than this value, on the 0 to 1 scale. An evaluation with no human score never matches. A value outside 0 to 1 raises ValueError.
​
float
default:"None"
Keyword-only. Keep evaluations whose human review score is less than or equal to this value, on the 0 to 1 scale. An evaluation with no human score never matches. A value outside 0 to 1 raises ValueError.
To match one exact score, pass the same value as both inclusive bounds: score_gte=0.5, score_lte=0.5 for the judge score, or human_score_gte=0.5, human_score_lte=0.5 for the human score.
At least one of the scoping filters (id, session_id, version_id, test_case_id, dataset_id, metric_id, run_id) described above must be passed to the method. status, user_id and the score bounds are additional filters and do not satisfy this requirement on their own.