Test fixture
Bug fixes, edge conditions, and minimal patch accuracy.
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 diagnosis, fix, tests. A pagination helper drops the final item when total is exactly divisible by page size because it subtracts one before computing the last page. Identify the bug and fix.
{
"metrics": {
"repro": {
"checks": [
{
"contains": [
"exactly divisible"
]
},
{
"contains": [
"final item"
]
},
{
"contains": [
"page size"
]
},
{
"contains": [
"last page"
]
}
]
},
"hidden": {
"checks": [
{
"contains": [
"Math.ceil"
]
},
{
"contains": [
"zero based"
]
},
{
"contains": [
"boundary test"
]
}
]
},
"diagnose": {
"checks": [
{
"contains": [
"off by one"
]
},
{
"contains": [
"pagination"
]
},
{
"contains": [
"subtracts one"
]
}
]
}
}
}temperature
0
max_tokens
420
timeout (s)
120
type
scored
file
debug-off-by-one-pagination-v2.json