{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:4HMJG45RTUDCFKZ4OEYHLHSWBT","short_pith_number":"pith:4HMJG45R","canonical_record":{"source":{"id":"2304.06528","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2023-04-13T13:29:01Z","cross_cats_sorted":[],"title_canon_sha256":"0ef07879454f6758976cb9d61e07af494b489fc93e0602c2f013465a8359502b","abstract_canon_sha256":"ba317ffffa9ff35e026952337c829558ff208f3e6193ab17eb481a6d690d41f4"},"schema_version":"1.0"},"canonical_sha256":"e1d89373b19d0622ab3c7130759e560cfff46454cd542217d8f6211750f6c2bf","source":{"kind":"arxiv","id":"2304.06528","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2304.06528","created_at":"2026-07-05T06:00:44Z"},{"alias_kind":"arxiv_version","alias_value":"2304.06528v1","created_at":"2026-07-05T06:00:44Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2304.06528","created_at":"2026-07-05T06:00:44Z"},{"alias_kind":"pith_short_12","alias_value":"4HMJG45RTUDC","created_at":"2026-07-05T06:00:44Z"},{"alias_kind":"pith_short_16","alias_value":"4HMJG45RTUDCFKZ4","created_at":"2026-07-05T06:00:44Z"},{"alias_kind":"pith_short_8","alias_value":"4HMJG45R","created_at":"2026-07-05T06:00:44Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:4HMJG45RTUDCFKZ4OEYHLHSWBT","target":"record","payload":{"canonical_record":{"source":{"id":"2304.06528","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2023-04-13T13:29:01Z","cross_cats_sorted":[],"title_canon_sha256":"0ef07879454f6758976cb9d61e07af494b489fc93e0602c2f013465a8359502b","abstract_canon_sha256":"ba317ffffa9ff35e026952337c829558ff208f3e6193ab17eb481a6d690d41f4"},"schema_version":"1.0"},"canonical_sha256":"e1d89373b19d0622ab3c7130759e560cfff46454cd542217d8f6211750f6c2bf","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:00:44.896936Z","signature_b64":"Su7KnlRA9oYwNnRVJqlpGDLY7b27f1HALVQcrjs37no2HGr63bhJq4yTfk3zK4I2vDvZyPGL7KJaY7P5wiHjDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e1d89373b19d0622ab3c7130759e560cfff46454cd542217d8f6211750f6c2bf","last_reissued_at":"2026-07-05T06:00:44.896435Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:00:44.896435Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2304.06528","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T06:00:44Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4EDfKtdAZeo0YQDy8DQzasEyI5a+eX16D1EjMsFNL/khbgLjjJsaY10JaIXuGpn6rlQboku5N/s1njY8+hHjAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T09:19:08.292857Z"},"content_sha256":"c49ab60eb90cd0b993c57051391d2376010bb1335be13270ff21b3ade73f21ae","schema_version":"1.0","event_id":"sha256:c49ab60eb90cd0b993c57051391d2376010bb1335be13270ff21b3ade73f21ae"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:4HMJG45RTUDCFKZ4OEYHLHSWBT","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Power-seeking can be probable and predictive for trained agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Janos Kramar, Victoria Krakovna","submitted_at":"2023-04-13T13:29:01Z","abstract_excerpt":"Power-seeking behavior is a key source of risk from advanced AI, but our theoretical understanding of this phenomenon is relatively limited. Building on existing theoretical results demonstrating power-seeking incentives for most reward functions, we investigate how the training process affects power-seeking incentives and show that they are still likely to hold for trained agents under some simplifying assumptions. We formally define the training-compatible goal set (the set of goals consistent with the training rewards) and assume that the trained agent learns a goal from this set. In a sett"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2304.06528","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/2304.06528/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":"2026-07-05T06:00:44Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"45+lfzyVtKbNyzi38zAim0iy3qFuDBGSEqrqlvUWLRuRURfAVUf0SK6mnzpMM6+Oh/2mVqnsNWRq/jGyInNvBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T09:19:08.293886Z"},"content_sha256":"c93020951d275a6efa28bdd91b277460d735158f1c94e9e55bc7905c88692093","schema_version":"1.0","event_id":"sha256:c93020951d275a6efa28bdd91b277460d735158f1c94e9e55bc7905c88692093"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/4HMJG45RTUDCFKZ4OEYHLHSWBT/bundle.json","state_url":"https://pith.science/pith/4HMJG45RTUDCFKZ4OEYHLHSWBT/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/4HMJG45RTUDCFKZ4OEYHLHSWBT/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":"2026-08-08T09:19:08Z","links":{"resolver":"https://pith.science/pith/4HMJG45RTUDCFKZ4OEYHLHSWBT","bundle":"https://pith.science/pith/4HMJG45RTUDCFKZ4OEYHLHSWBT/bundle.json","state":"https://pith.science/pith/4HMJG45RTUDCFKZ4OEYHLHSWBT/state.json","well_known_bundle":"https://pith.science/.well-known/pith/4HMJG45RTUDCFKZ4OEYHLHSWBT/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:4HMJG45RTUDCFKZ4OEYHLHSWBT","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":"ba317ffffa9ff35e026952337c829558ff208f3e6193ab17eb481a6d690d41f4","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2023-04-13T13:29:01Z","title_canon_sha256":"0ef07879454f6758976cb9d61e07af494b489fc93e0602c2f013465a8359502b"},"schema_version":"1.0","source":{"id":"2304.06528","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2304.06528","created_at":"2026-07-05T06:00:44Z"},{"alias_kind":"arxiv_version","alias_value":"2304.06528v1","created_at":"2026-07-05T06:00:44Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2304.06528","created_at":"2026-07-05T06:00:44Z"},{"alias_kind":"pith_short_12","alias_value":"4HMJG45RTUDC","created_at":"2026-07-05T06:00:44Z"},{"alias_kind":"pith_short_16","alias_value":"4HMJG45RTUDCFKZ4","created_at":"2026-07-05T06:00:44Z"},{"alias_kind":"pith_short_8","alias_value":"4HMJG45R","created_at":"2026-07-05T06:00:44Z"}],"graph_snapshots":[{"event_id":"sha256:c93020951d275a6efa28bdd91b277460d735158f1c94e9e55bc7905c88692093","target":"graph","created_at":"2026-07-05T06:00:44Z","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/2304.06528/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Power-seeking behavior is a key source of risk from advanced AI, but our theoretical understanding of this phenomenon is relatively limited. Building on existing theoretical results demonstrating power-seeking incentives for most reward functions, we investigate how the training process affects power-seeking incentives and show that they are still likely to hold for trained agents under some simplifying assumptions. We formally define the training-compatible goal set (the set of goals consistent with the training rewards) and assume that the trained agent learns a goal from this set. In a sett","authors_text":"Janos Kramar, Victoria Krakovna","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2023-04-13T13:29:01Z","title":"Power-seeking can be probable and predictive for trained agents"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2304.06528","kind":"arxiv","version":1},"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:c49ab60eb90cd0b993c57051391d2376010bb1335be13270ff21b3ade73f21ae","target":"record","created_at":"2026-07-05T06:00:44Z","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":"ba317ffffa9ff35e026952337c829558ff208f3e6193ab17eb481a6d690d41f4","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2023-04-13T13:29:01Z","title_canon_sha256":"0ef07879454f6758976cb9d61e07af494b489fc93e0602c2f013465a8359502b"},"schema_version":"1.0","source":{"id":"2304.06528","kind":"arxiv","version":1}},"canonical_sha256":"e1d89373b19d0622ab3c7130759e560cfff46454cd542217d8f6211750f6c2bf","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e1d89373b19d0622ab3c7130759e560cfff46454cd542217d8f6211750f6c2bf","first_computed_at":"2026-07-05T06:00:44.896435Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T06:00:44.896435Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Su7KnlRA9oYwNnRVJqlpGDLY7b27f1HALVQcrjs37no2HGr63bhJq4yTfk3zK4I2vDvZyPGL7KJaY7P5wiHjDg==","signature_status":"signed_v1","signed_at":"2026-07-05T06:00:44.896936Z","signed_message":"canonical_sha256_bytes"},"source_id":"2304.06528","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c49ab60eb90cd0b993c57051391d2376010bb1335be13270ff21b3ade73f21ae","sha256:c93020951d275a6efa28bdd91b277460d735158f1c94e9e55bc7905c88692093"],"state_sha256":"3c5c53de8d91f137d8c3ea49aa6ae85a39a7e22825c86f7037a6c27df12a83fc"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xOFSnWELYIRZPrfgB74CKyPENAivmVWWg4juzTO0TxaV3u1YuFWjnAR0fU8PpD81W/0vSuk+4GMHS18QdQdABg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T09:19:08.301202Z","bundle_sha256":"2706b5245db904b8c6f7438f0d927d82eac4095569f6b17b7e01025968c66a1d"}}