{"schema_version":"papers/paper-detail-v1","title":"Hack-Verifiable Environments: Towards Evaluating Reward Hacking at Scale","surface":"papers","opportunity_kernel":{"paper_id":"4acc2ead-9c36-4ff0-98ae-715e55052278","title":"Hack-Verifiable Environments: Towards Evaluating Reward Hacking at Scale","authors":["Amit Roth","Ankur Samanta","Matan Halevy","Yoav Levine","Yonathan Efroni"],"arxiv_id":"2605.20744v1","doi":null,"published_at":"2026-05-20T05:46:52.000Z","score_object":{"overall":{"value":7,"scale":"0-10","confidence":0.85,"confidence_reason":"Backfilled from persisted papers.viability_score.","model_version":"phase0-backfill-v1","pipeline_version":"phase0-kernel-v1","computed_at":"2026-05-21T20:11:32.870Z","fresh_until":"2026-06-20T20:11:32.870Z","is_stale":false,"source_count":1,"missingness":[]},"technical":{"value":1.4,"scale":"0-10","confidence":0.55,"confidence_reason":"Backfilled from paper_extraction_scorecards.reconstruction_score.","model_version":"phase0-backfill-v1","pipeline_version":"phase0-kernel-v1","computed_at":"2026-05-21T20:13:52.322Z","fresh_until":"2026-06-04T20:13:52.322Z","is_stale":false,"source_count":1,"missingness":["reproducibility_results.reproducibility_score","deployability_scores.score"]},"commercial":{"value":5,"scale":"0-10","confidence":0.75,"confidence_reason":"Backfilled from persisted commercial_flags and repo availability.","model_version":"phase0-backfill-v1","pipeline_version":"phase0-kernel-v1","computed_at":"2026-05-21T20:11:32.870Z","fresh_until":"2026-06-20T20:11:32.870Z","is_stale":false,"source_count":2,"missingness":[]},"market":{"value":5.5,"scale":"0-10","confidence":0.45,"confidence_reason":"Heuristic market score materialized from paper metadata, repo availability, and deployability.","model_version":"phase0-backfill-v1","pipeline_version":"phase0-kernel-v1","computed_at":"2026-05-21T20:29:42.521Z","fresh_until":"2026-06-04T20:29:42.521Z","is_stale":false,"source_count":1,"missingness":[]},"team":{"value":4.15,"scale":"0-10","confidence":0.28,"confidence_reason":"Heuristic fallback from paper author count.","model_version":"phase0-backfill-v1","pipeline_version":"phase0-kernel-v1","computed_at":"2026-05-21T20:11:32.870Z","fresh_until":"2026-06-04T20:11:32.870Z","is_stale":false,"source_count":1,"missingness":["paper_author_affiliations"]},"methodology":{"value":4.7,"scale":"0-10","confidence":0.82,"confidence_reason":"Backfilled from paper_extraction_scorecards.total_score.","model_version":"phase0-backfill-v1","pipeline_version":"phase0-kernel-v1","computed_at":"2026-05-21T20:13:52.322Z","fresh_until":"2026-06-20T20:13:52.322Z","is_stale":false,"source_count":2,"missingness":[]}},"evidence_receipt":{"freshness":"stale","proof_status":"unverified","repo_status":"active","references_count":0,"source_count":4,"coverage":0.6667,"missingness":["references","proof_status"],"unresolved_unknowns":["proof verification has not been recorded yet"],"last_verification_at":"2026-05-21T20:29:42.521Z"},"lineage_hash":"0193eb600059e66f6373ec5ec01ca927f13af738f1f07e774416e2b5d114553b"},"distribution":null,"replication_evidence":[],"author_dna":[]}