Test fixture
Grounded answers under adversarial or missing-context prompts.
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, supported_facts, uncertainty. A pagination API returns nextCursor only when more rows exist. For exactly pageSize rows and no extra row, decide if nextCursor should be present.
{
"metrics": {
"accuracy": {
"checks": [
{
"contains": [
"no extra row"
]
},
{
"contains": [
"nextCursor"
]
},
{
"contains": [
"more rows"
]
}
]
},
"fabricated": {
"checks": [
{
"contains": [
"exactly pageSize"
]
},
{
"contains": [
"edge case"
]
},
{
"contains": [
"do not fabricate"
]
}
]
}
}
}temperature
0
max_tokens
460
timeout (s)
120
type
scored
file
halluc-edge-pagination.json