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. A migration must notify users 7 days before billing change, cannot run on weekends, and billing changes Monday May 20. Latest valid migration date?
{
"metrics": {
"accuracy": {
"checks": [
{
"contains": [
"May 13"
]
},
{
"contains": [
"7 days"
]
},
{
"contains": [
"Monday May 20"
]
}
]
},
"evidence": {
"checks": [
{
"contains": [
"weekends"
]
},
{
"contains": [
"notify users"
]
},
{
"contains": [
"latest"
]
}
]
},
"constraint": {
"checks": [
{
"contains": [
"constraint"
]
},
{
"contains": [
"billing"
]
},
{
"contains": [
"date"
]
}
]
}
}
}temperature
0
max_tokens
500
timeout (s)
120
type
scored
file
reason-constraint-subscription-migration.json