Datasets:
File size: 3,000 Bytes
aabcff2 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 | {
"task_id": "UMA_GRADE_A_CONDITIONAL_CHALLENGE_V1",
"task_type": "four_action_protocol_outcome_decision",
"estimand": "P(protocol later rejects proposal | actually challenged, Grade-A linked, decision-time evidence)",
"action_space": [
"Accept",
"Investigate",
"Challenge",
"Abstain"
],
"decision_time": "Polygon OOV2 ProposePrice block timestamp",
"deadline": "OOV2 proposal expiration_time",
"label": "proposal_rejected_by_protocol",
"ground_truth_status": "protocol_resolution_only",
"selection_warning": "Outcomes are observed only for actually challenged requests; do not generalize causal policy value to undisputed requests.",
"model_features": [
"log_bond_raw",
"log_final_fee_raw",
"log_reward_plus1_raw",
"reward_to_bond_ratio",
"liveness_hours",
"proposal_latency_hours",
"ancillary_bytes",
"requester_is_polymarket_adapter",
"proposer_prior_completed",
"proposer_prior_dispute_rate",
"proposer_prior_rejection_rate",
"requester_prior_completed",
"requester_prior_dispute_rate",
"requester_prior_rejection_rate",
"proposed_price_class",
"adapter_version"
],
"forbidden_model_fields": [
"dispute_tx",
"dispute_time",
"resolved_price_raw",
"gross_payout_raw",
"settlement_tx",
"settlement_time",
"terminal_outcome",
"dvm_request_id",
"dvm_resolved_price_raw",
"proposal_rejected_by_protocol"
],
"action_thresholds": {
"Accept": "p<=0.20",
"Investigate": "0.20<p<0.40 or 0.60<=p<0.80",
"Abstain": "0.40<=p<0.60",
"Challenge": "p>=0.80"
},
"nominal_cost_matrix": {
"Accept": {
"upheld": 0.0,
"rejected": 1.0
},
"Challenge": {
"upheld": 1.0,
"rejected": 0.0
},
"Investigate": {
"upheld": 0.1,
"rejected": 0.1
},
"Abstain": {
"upheld": 0.2,
"rejected": 0.2
}
},
"cost_warning": "Normalized scenario costs are pre-registered evaluation assumptions, not observed token or USD payoffs.",
"evaluation_extensions": {
"claim_level_faithfulness": "data/applications/trustworthy_ai_requirements_audit/evidence_faithfulness.parquet",
"timestamped_usd_regret": "data/applications/trustworthy_ai_usd_economics/economic_regret_scenarios.parquet",
"independent_ground_truth": "data/applications/trustworthy_ai_independent_truth/independent_ground_truth.parquet",
"cross_protocol_fairness": "data/applications/trustworthy_ai_cross_protocol_fairness/cross_protocol_metrics.parquet",
"complete_test_panel": "data/applications/trustworthy_ai_complete_task/complete_test_decision_panel.parquet"
},
"independent_truth_policy": "Retrospective external labels are evaluation-only and never model inputs.",
"usd_cost_policy": "Gas/payoff FX is measured historically; investigation cost and capital APR remain registered scenarios.",
"split": "strict chronological 60/20/20",
"seeds": [
7,
17,
27,
37,
47
]
}
|