EvalArtifacts

Final outputs of an eval run.

Attributes

attributesummaryEvalSummary

attributesummary_strict_primaryEvalSummary | None

Optional strict variant (e.g. primary-only). None for plain model evals.

attributeper_row_resultslist[dict[str, Any]]

Raw per-row, per-query results so downstream tooling can re-score.

func__init__(self, summary, summary_strict_primary, per_row_results) -> None

paramself

paramsummaryEvalSummary

paramsummary_strict_primaryEvalSummary | None

paramper_row_resultslist[dict[str, Any]]

Returns

None