amana / eval /testset.json
Misbahuddin's picture
Phase 4: eval harness + CI (deterministic safety gate + LLM-judge)
9561b34
Raw
History Blame Contribute Delete
3.28 kB
[
{
"id": "camp-001",
"path": "data/campaigns/camp-001.json",
"expected": {
"recommendation": "APPROVE",
"primary_rules": []
}
},
{
"id": "camp-002",
"path": "data/campaigns/camp-002.json",
"expected": {
"recommendation": "APPROVE",
"primary_rules": []
}
},
{
"id": "camp-003",
"path": "data/campaigns/camp-003.json",
"expected": {
"recommendation": "APPROVE",
"primary_rules": []
}
},
{
"id": "camp-004",
"path": "data/campaigns/camp-004.json",
"expected": {
"recommendation": "APPROVE",
"primary_rules": []
}
},
{
"id": "camp-005",
"path": "data/campaigns/camp-005.json",
"expected": {
"recommendation": "REJECT",
"primary_rules": [
"PROH-3",
"PROH-7"
]
}
},
{
"id": "camp-006",
"path": "data/campaigns/camp-006.json",
"expected": {
"recommendation": "REJECT",
"primary_rules": [
"PROH-4"
]
}
},
{
"id": "camp-007",
"path": "data/campaigns/camp-007.json",
"expected": {
"recommendation": "REJECT",
"primary_rules": [
"COMP-3"
]
}
},
{
"id": "camp-008",
"path": "data/campaigns/camp-008.json",
"expected": {
"recommendation": "REJECT",
"primary_rules": [
"PROH-2"
]
}
},
{
"id": "camp-009",
"path": "data/campaigns/camp-009.json",
"expected": {
"recommendation": "ESCALATE",
"primary_rules": [
"COMP-1",
"COMP-4"
]
}
},
{
"id": "camp-010",
"path": "data/campaigns/camp-010.json",
"expected": {
"recommendation": "ESCALATE",
"primary_rules": [
"COMP-2"
]
}
},
{
"id": "camp-011",
"path": "data/campaigns/camp-011.json",
"expected": {
"recommendation": "ESCALATE",
"primary_rules": [
"ELIG-2",
"ELIG-4"
]
}
},
{
"id": "camp-012",
"path": "data/campaigns/camp-012.json",
"expected": {
"recommendation": "ESCALATE",
"primary_rules": [
"ELIG-4"
]
}
},
{
"id": "camp-013",
"path": "data/campaigns/camp-013.json",
"expected": {
"recommendation": "ESCALATE",
"primary_rules": [
"CONT-3",
"ELIG-2"
]
}
},
{
"id": "camp-014",
"path": "data/campaigns/camp-014.json",
"expected": {
"recommendation": "ESCALATE",
"primary_rules": [
"CONT-2"
]
}
},
{
"id": "camp-015",
"path": "data/campaigns/camp-015.json",
"expected": {
"recommendation": "ESCALATE",
"primary_rules": [
"DEC-6",
"ELIG-2"
]
}
},
{
"id": "camp-016",
"path": "data/campaigns/camp-016.json",
"expected": {
"recommendation": "ESCALATE",
"primary_rules": [
"CONT-5",
"ELIG-2"
]
}
},
{
"id": "camp-017",
"path": "data/campaigns/camp-017.json",
"expected": {
"recommendation": "APPROVE",
"primary_rules": []
}
},
{
"id": "camp-018",
"path": "data/campaigns/camp-018.json",
"expected": {
"recommendation": "APPROVE",
"primary_rules": []
}
}
]