{# Evals > one run: what it ran on and how it went, then each case, failed first and opened: what it asked and answered, each check by its rule, its calls; what it took and cost. Compare goes to the run of the same dataset before it. #} {% from "components/macros/layout.html" import card, facts as fact_list, goto %} {% from "components/macros/table.html" import data_table %} {% from "pages/overseer/ai/_evals_macros.html" import case_result %} {% macro result(row) %}{{ case_result(row.result) }}{% endmacro %} {% call card("Run " ~ run.id, id="ai-evals-run") %} {{ fact_list(facts) }} {% if compare %}

{{ goto(compare, "Compare with the run before", "data-evals-compare") }}

{% endif %} {% endcall %}
{% call card(None, id="ai-evals-cases-card") %} {{ data_table([ {"key": "case", "label": "Case", "kind": "link", "wrap": True}, {"key": "result_cell", "label": "Result", "kind": "status"}, {"key": "took", "label": "Took", "align": "right"}, {"key": "cost", "label": "Cost", "align": "right"}, ], cases, empty="This run kept no cases", detail=result) }} {% endcall %}