{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:TMQYUJLVSFOTOY7BF4SBVCLE3H","short_pith_number":"pith:TMQYUJLV","schema_version":"1.0","canonical_sha256":"9b218a2575915d3763e12f241a8964d9e9f4f45121078606b7defffe5879c306","source":{"kind":"arxiv","id":"2605.20744","version":1},"attestation_state":"computed","paper":{"title":"Hack-Verifiable Environments: Towards Evaluating Reward Hacking at Scale","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Amit Roth, Ankur Samanta, Matan Halevy, Yoav Levine, Yonathan Efroni","submitted_at":"2026-05-20T05:46:52Z","abstract_excerpt":"Aligning autonomous agents with human intent remains a central challenge in modern AI. A key manifestation of this challenge is reward hacking, whereby agents appear successful under the evaluation signal while violating the intended objective. Reward hacking has been observed across a wide range of settings, yet methods for reliably measuring it at scale remain lacking. In this work, we introduce a new evaluation paradigm for measuring reward hacking. Whereas prior studies have primarily analyzed it post hoc by inspecting agent trajectories, we instead embed detectable reward hacking opportun"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2605.20744","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-20T05:46:52Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ff5952a0782a4f82d008f6b9aa8b23f902160f2a4cf95d62571be7febb17754b","abstract_canon_sha256":"ccf27647ca044b771adc0168f87d893d0e04b23eaff44601efae0d2ea2b80720"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T01:04:51.881903Z","signature_b64":"CYmHEX1hz3d5oMnBRDC63UGf5M44O1XbRuozZHrvcmbc/i0oxMPMyZLTIrEyGqGq9PAy3MWgOgdnMUE3axVBDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9b218a2575915d3763e12f241a8964d9e9f4f45121078606b7defffe5879c306","last_reissued_at":"2026-05-21T01:04:51.881137Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T01:04:51.881137Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Hack-Verifiable Environments: Towards Evaluating Reward Hacking at Scale","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Amit Roth, Ankur Samanta, Matan Halevy, Yoav Levine, Yonathan Efroni","submitted_at":"2026-05-20T05:46:52Z","abstract_excerpt":"Aligning autonomous agents with human intent remains a central challenge in modern AI. A key manifestation of this challenge is reward hacking, whereby agents appear successful under the evaluation signal while violating the intended objective. Reward hacking has been observed across a wide range of settings, yet methods for reliably measuring it at scale remain lacking. In this work, we introduce a new evaluation paradigm for measuring reward hacking. Whereas prior studies have primarily analyzed it post hoc by inspecting agent trajectories, we instead embed detectable reward hacking opportun"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.20744","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.20744/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2605.20744","created_at":"2026-05-21T01:04:51.881266+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.20744v1","created_at":"2026-05-21T01:04:51.881266+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.20744","created_at":"2026-05-21T01:04:51.881266+00:00"},{"alias_kind":"pith_short_12","alias_value":"TMQYUJLVSFOT","created_at":"2026-05-21T01:04:51.881266+00:00"},{"alias_kind":"pith_short_16","alias_value":"TMQYUJLVSFOTOY7B","created_at":"2026-05-21T01:04:51.881266+00:00"},{"alias_kind":"pith_short_8","alias_value":"TMQYUJLV","created_at":"2026-05-21T01:04:51.881266+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TMQYUJLVSFOTOY7BF4SBVCLE3H","json":"https://pith.science/pith/TMQYUJLVSFOTOY7BF4SBVCLE3H.json","graph_json":"https://pith.science/api/pith-number/TMQYUJLVSFOTOY7BF4SBVCLE3H/graph.json","events_json":"https://pith.science/api/pith-number/TMQYUJLVSFOTOY7BF4SBVCLE3H/events.json","paper":"https://pith.science/paper/TMQYUJLV"},"agent_actions":{"view_html":"https://pith.science/pith/TMQYUJLVSFOTOY7BF4SBVCLE3H","download_json":"https://pith.science/pith/TMQYUJLVSFOTOY7BF4SBVCLE3H.json","view_paper":"https://pith.science/paper/TMQYUJLV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.20744&json=true","fetch_graph":"https://pith.science/api/pith-number/TMQYUJLVSFOTOY7BF4SBVCLE3H/graph.json","fetch_events":"https://pith.science/api/pith-number/TMQYUJLVSFOTOY7BF4SBVCLE3H/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TMQYUJLVSFOTOY7BF4SBVCLE3H/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TMQYUJLVSFOTOY7BF4SBVCLE3H/action/storage_attestation","attest_author":"https://pith.science/pith/TMQYUJLVSFOTOY7BF4SBVCLE3H/action/author_attestation","sign_citation":"https://pith.science/pith/TMQYUJLVSFOTOY7BF4SBVCLE3H/action/citation_signature","submit_replication":"https://pith.science/pith/TMQYUJLVSFOTOY7BF4SBVCLE3H/action/replication_record"}},"created_at":"2026-05-21T01:04:51.881266+00:00","updated_at":"2026-05-21T01:04:51.881266+00:00"}