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 payment API claims retries are idempotent because the client sends the same amount each time. Show why that is insufficient without an idempotency key.
{
"metrics": {
"accuracy": {
"checks": [
{
"contains": [
"same amount"
]
},
{
"contains": [
"not enough"
]
},
{
"contains": [
"idempotency key"
]
}
]
},
"evidence": {
"checks": [
{
"contains": [
"payment API"
]
},
{
"contains": [
"retry"
]
},
{
"contains": [
"duplicate charge"
]
}
]
},
"constraint": {
"checks": [
{
"contains": [
"counterexample"
]
},
{
"contains": [
"request identity"
]
},
{
"contains": [
"side effect"
]
}
]
}
}
}temperature
0
max_tokens
560
timeout (s)
120
type
scored
file
reasoning-fce-idempotent-retry-proof.json