Test fixture
Arithmetic, symbolic steps, and structured problem solving.
The model receives the prompt (and optional system message). The run uses scorer rubric_json_metrics with the JSON configuration below. Pass/fail and partial credit are determined entirely by that scorer against the model output; no human grading.
Return JSON only with keys answer, evidence, constraints. Migration doc says backfill completed at 10:00. Query logs show null new_user_id at 10:15 for migrated rows. Identify implication and safest action.
{
"metrics": {
"accuracy": {
"checks": [
{
"contains": [
"10:15"
]
},
{
"contains": [
"null"
]
},
{
"contains": [
"backfill"
]
}
]
},
"evidence": {
"checks": [
{
"contains": [
"10:00"
]
},
{
"contains": [
"migration doc"
]
},
{
"contains": [
"safest action"
]
}
]
},
"constraint": {
"checks": [
{
"contains": [
"multi-artifact"
]
},
{
"contains": [
"rollback"
]
},
{
"contains": [
"verify"
]
}
]
}
}
}temperature
0
max_tokens
500
timeout (s)
120
type
scored
file
reasoning-mas-schema-backfill.json