{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:QMYUDFWAX4FXXBHUYD2P2UWX3Z","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"d5c02fa9d9f6cf323e0ad8233d295fd38a3969fc9140e7f20bc933ea4299d83c","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-28T06:37:10Z","title_canon_sha256":"4a99ca29dd93783ca280883a4e2b876f7f49751323e1ebdaf84df9f4d166d29f"},"schema_version":"1.0","source":{"id":"2506.22777","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.22777","created_at":"2026-07-05T11:36:33Z"},{"alias_kind":"arxiv_version","alias_value":"2506.22777v2","created_at":"2026-07-05T11:36:33Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.22777","created_at":"2026-07-05T11:36:33Z"},{"alias_kind":"pith_short_12","alias_value":"QMYUDFWAX4FX","created_at":"2026-07-05T11:36:33Z"},{"alias_kind":"pith_short_16","alias_value":"QMYUDFWAX4FXXBHU","created_at":"2026-07-05T11:36:33Z"},{"alias_kind":"pith_short_8","alias_value":"QMYUDFWA","created_at":"2026-07-05T11:36:33Z"}],"graph_snapshots":[{"event_id":"sha256:5c9e7168dcd189334d3b8557cbe88dc9556c72113d38febaa9243b68eb375602","target":"graph","created_at":"2026-07-05T11:36:33Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2506.22777/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Language models trained with reinforcement learning (RL) can engage in reward hacking--the exploitation of unintended strategies for high reward--without revealing this behavior in their chain-of-thought reasoning. This makes the detection of reward hacking difficult, posing risks for high-stakes applications. We propose verbalization fine-tuning (VFT), a pre-RL fine-tuning intervention that trains models to explicitly acknowledge when they are influenced by prompt cues--hints which point to incorrect answers (e.g., \"a Stanford professor thinks the answer is A\"). To evaluate VFT, we subsequent","authors_text":"Andy Arditi, Joe Benton, Julian Michael, Marvin Li, Miles Turpin","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-28T06:37:10Z","title":"Teaching Models to Verbalize Reward Hacking in Chain-of-Thought Reasoning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.22777","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:c7352de03681aaa719e8a5c43df86f28f90d8de721c800c3e6c5a53bfb05c1c1","target":"record","created_at":"2026-07-05T11:36:33Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"d5c02fa9d9f6cf323e0ad8233d295fd38a3969fc9140e7f20bc933ea4299d83c","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-28T06:37:10Z","title_canon_sha256":"4a99ca29dd93783ca280883a4e2b876f7f49751323e1ebdaf84df9f4d166d29f"},"schema_version":"1.0","source":{"id":"2506.22777","kind":"arxiv","version":2}},"canonical_sha256":"83314196c0bf0b7b84f4c0f4fd52d7de5c2fecc470f21faa21bd1108d00c7560","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"83314196c0bf0b7b84f4c0f4fd52d7de5c2fecc470f21faa21bd1108d00c7560","first_computed_at":"2026-07-05T11:36:33.745862Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:36:33.745862Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"y4TPQFHeHVzUZw37r5dgibNgzOBXus6IZxQCbTewQ2voQJcXDF24yvKWqF1+kBQmM32amBjHpb6gzjzHN2TLDA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:36:33.746376Z","signed_message":"canonical_sha256_bytes"},"source_id":"2506.22777","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c7352de03681aaa719e8a5c43df86f28f90d8de721c800c3e6c5a53bfb05c1c1","sha256:5c9e7168dcd189334d3b8557cbe88dc9556c72113d38febaa9243b68eb375602"],"state_sha256":"a9684a05b61adfc581894e547fc6e00e8e6692b65d8015e1977d1996b4d53771"}