{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:NYHYQYTDMNFPZCE3N6TKET2FS3","short_pith_number":"pith:NYHYQYTD","canonical_record":{"source":{"id":"2502.13295","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-02-18T21:32:24Z","cross_cats_sorted":[],"title_canon_sha256":"cffdfdb79869fe4d25496b2abe5789bbbbdddca143146a8c34f93b27d5734606","abstract_canon_sha256":"e22f45aa9845c4874fe3249401a1affb4bedbfdc8ee3a6f7949034af64cdabb1"},"schema_version":"1.0"},"canonical_sha256":"6e0f886263634afc889b6fa6a24f4596cac57cbbd740912c2a6e95ade0093d09","source":{"kind":"arxiv","id":"2502.13295","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.13295","created_at":"2026-07-05T11:59:57Z"},{"alias_kind":"arxiv_version","alias_value":"2502.13295v3","created_at":"2026-07-05T11:59:57Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.13295","created_at":"2026-07-05T11:59:57Z"},{"alias_kind":"pith_short_12","alias_value":"NYHYQYTDMNFP","created_at":"2026-07-05T11:59:57Z"},{"alias_kind":"pith_short_16","alias_value":"NYHYQYTDMNFPZCE3","created_at":"2026-07-05T11:59:57Z"},{"alias_kind":"pith_short_8","alias_value":"NYHYQYTD","created_at":"2026-07-05T11:59:57Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:NYHYQYTDMNFPZCE3N6TKET2FS3","target":"record","payload":{"canonical_record":{"source":{"id":"2502.13295","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-02-18T21:32:24Z","cross_cats_sorted":[],"title_canon_sha256":"cffdfdb79869fe4d25496b2abe5789bbbbdddca143146a8c34f93b27d5734606","abstract_canon_sha256":"e22f45aa9845c4874fe3249401a1affb4bedbfdc8ee3a6f7949034af64cdabb1"},"schema_version":"1.0"},"canonical_sha256":"6e0f886263634afc889b6fa6a24f4596cac57cbbd740912c2a6e95ade0093d09","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:59:57.472457Z","signature_b64":"AG3apX/Lnkxy/jlx1u8Q7UhIrXkHdoJT1u3nOTSBoLB5U5n91CK0FtckMMNeVUCx0TDe1fR505LNsITta0fPBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6e0f886263634afc889b6fa6a24f4596cac57cbbd740912c2a6e95ade0093d09","last_reissued_at":"2026-07-05T11:59:57.471964Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:59:57.471964Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2502.13295","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:59:57Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Pdw41461wS0ZsBDL3NwJIOO6i947id9zo+mf3KaWxaDjnrVdYb6FBttLBdoJNvqv0+dp6DMfqofCCKfYgveTBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T09:54:20.198073Z"},"content_sha256":"714f41a82e8765d461283940e13da721a883fddf382f293cb6534d8660caadb7","schema_version":"1.0","event_id":"sha256:714f41a82e8765d461283940e13da721a883fddf382f293cb6534d8660caadb7"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:NYHYQYTDMNFPZCE3N6TKET2FS3","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Demonstrating specification gaming in reasoning models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Alexander Bondarenko, Denis Volk, Dmitrii Volkov, Jeffrey Ladish","submitted_at":"2025-02-18T21:32:24Z","abstract_excerpt":"We demonstrate LLM agent specification gaming by instructing models to win against a chess engine. We find reasoning models like OpenAI o3 and DeepSeek R1 will often hack the benchmark by default, while language models like GPT-4o and Claude 3.5 Sonnet need to be told that normal play won't work to hack.\n  We improve upon prior work like (Hubinger et al., 2024; Meinke et al., 2024; Weij et al., 2024) by using realistic task prompts and avoiding excess nudging. Our results suggest reasoning models may resort to hacking to solve difficult problems, as observed in OpenAI (2024)'s o1 Docker escape"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.13295","kind":"arxiv","version":3},"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/2502.13295/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:59:57Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"W/XSzNXnz75UdYiYJfq91ujtPyzp3nX0r3yU3jX9Nzb8GVeK0Vpb2ZjiGSZZPSvFaBqqKbiETamsT4kbmk43CQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T09:54:20.198807Z"},"content_sha256":"4edf13f0f0e8d3acd904d61fc5dbfefb29de7868884658b3952488753784fb4a","schema_version":"1.0","event_id":"sha256:4edf13f0f0e8d3acd904d61fc5dbfefb29de7868884658b3952488753784fb4a"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/NYHYQYTDMNFPZCE3N6TKET2FS3/bundle.json","state_url":"https://pith.science/pith/NYHYQYTDMNFPZCE3N6TKET2FS3/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/NYHYQYTDMNFPZCE3N6TKET2FS3/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-05T09:54:20Z","links":{"resolver":"https://pith.science/pith/NYHYQYTDMNFPZCE3N6TKET2FS3","bundle":"https://pith.science/pith/NYHYQYTDMNFPZCE3N6TKET2FS3/bundle.json","state":"https://pith.science/pith/NYHYQYTDMNFPZCE3N6TKET2FS3/state.json","well_known_bundle":"https://pith.science/.well-known/pith/NYHYQYTDMNFPZCE3N6TKET2FS3/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:NYHYQYTDMNFPZCE3N6TKET2FS3","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":"e22f45aa9845c4874fe3249401a1affb4bedbfdc8ee3a6f7949034af64cdabb1","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-02-18T21:32:24Z","title_canon_sha256":"cffdfdb79869fe4d25496b2abe5789bbbbdddca143146a8c34f93b27d5734606"},"schema_version":"1.0","source":{"id":"2502.13295","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.13295","created_at":"2026-07-05T11:59:57Z"},{"alias_kind":"arxiv_version","alias_value":"2502.13295v3","created_at":"2026-07-05T11:59:57Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.13295","created_at":"2026-07-05T11:59:57Z"},{"alias_kind":"pith_short_12","alias_value":"NYHYQYTDMNFP","created_at":"2026-07-05T11:59:57Z"},{"alias_kind":"pith_short_16","alias_value":"NYHYQYTDMNFPZCE3","created_at":"2026-07-05T11:59:57Z"},{"alias_kind":"pith_short_8","alias_value":"NYHYQYTD","created_at":"2026-07-05T11:59:57Z"}],"graph_snapshots":[{"event_id":"sha256:4edf13f0f0e8d3acd904d61fc5dbfefb29de7868884658b3952488753784fb4a","target":"graph","created_at":"2026-07-05T11:59:57Z","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/2502.13295/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We demonstrate LLM agent specification gaming by instructing models to win against a chess engine. We find reasoning models like OpenAI o3 and DeepSeek R1 will often hack the benchmark by default, while language models like GPT-4o and Claude 3.5 Sonnet need to be told that normal play won't work to hack.\n  We improve upon prior work like (Hubinger et al., 2024; Meinke et al., 2024; Weij et al., 2024) by using realistic task prompts and avoiding excess nudging. Our results suggest reasoning models may resort to hacking to solve difficult problems, as observed in OpenAI (2024)'s o1 Docker escape","authors_text":"Alexander Bondarenko, Denis Volk, Dmitrii Volkov, Jeffrey Ladish","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-02-18T21:32:24Z","title":"Demonstrating specification gaming in reasoning models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.13295","kind":"arxiv","version":3},"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":"sha256:714f41a82e8765d461283940e13da721a883fddf382f293cb6534d8660caadb7","target":"record","created_at":"2026-07-05T11:59:57Z","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":"e22f45aa9845c4874fe3249401a1affb4bedbfdc8ee3a6f7949034af64cdabb1","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-02-18T21:32:24Z","title_canon_sha256":"cffdfdb79869fe4d25496b2abe5789bbbbdddca143146a8c34f93b27d5734606"},"schema_version":"1.0","source":{"id":"2502.13295","kind":"arxiv","version":3}},"canonical_sha256":"6e0f886263634afc889b6fa6a24f4596cac57cbbd740912c2a6e95ade0093d09","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"6e0f886263634afc889b6fa6a24f4596cac57cbbd740912c2a6e95ade0093d09","first_computed_at":"2026-07-05T11:59:57.471964Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:59:57.471964Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"AG3apX/Lnkxy/jlx1u8Q7UhIrXkHdoJT1u3nOTSBoLB5U5n91CK0FtckMMNeVUCx0TDe1fR505LNsITta0fPBA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:59:57.472457Z","signed_message":"canonical_sha256_bytes"},"source_id":"2502.13295","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:714f41a82e8765d461283940e13da721a883fddf382f293cb6534d8660caadb7","sha256:4edf13f0f0e8d3acd904d61fc5dbfefb29de7868884658b3952488753784fb4a"],"state_sha256":"fd0806274f8756e073c98ba71991115cbc8e906352732751819e51e1807f986b"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"wx/5d2rXx8QjvxIgO/n5qfR4R4mIFCo8rsixK1Z2fqHwHCUbWtLP40KmqaxA1/UfyuHSc1oXuo8pLv8EOb4JBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-05T09:54:20.209060Z","bundle_sha256":"05e6958ebcf7bcd0ff9c992371a4a5984f797fcd9e883cfcca541c98ebd533f0"}}