{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:YRKEMS2YFCGZ4Y3NVKD4QD7QDT","short_pith_number":"pith:YRKEMS2Y","canonical_record":{"source":{"id":"2408.15116","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-08-27T14:55:15Z","cross_cats_sorted":[],"title_canon_sha256":"926e331176d900d0930fa9721d111e217d07a0fcc193f5a689910b8cd5d3f5d2","abstract_canon_sha256":"016bb668ae7163a5606a58208693ef6cd762090aa10277c64d701b35821cbfab"},"schema_version":"1.0"},"canonical_sha256":"c454464b58288d9e636daa87c80ff01ccb7657ed832370301a81bcef57f68626","source":{"kind":"arxiv","id":"2408.15116","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2408.15116","created_at":"2026-07-05T08:59:54Z"},{"alias_kind":"arxiv_version","alias_value":"2408.15116v1","created_at":"2026-07-05T08:59:54Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.15116","created_at":"2026-07-05T08:59:54Z"},{"alias_kind":"pith_short_12","alias_value":"YRKEMS2YFCGZ","created_at":"2026-07-05T08:59:54Z"},{"alias_kind":"pith_short_16","alias_value":"YRKEMS2YFCGZ4Y3N","created_at":"2026-07-05T08:59:54Z"},{"alias_kind":"pith_short_8","alias_value":"YRKEMS2Y","created_at":"2026-07-05T08:59:54Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:YRKEMS2YFCGZ4Y3NVKD4QD7QDT","target":"record","payload":{"canonical_record":{"source":{"id":"2408.15116","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-08-27T14:55:15Z","cross_cats_sorted":[],"title_canon_sha256":"926e331176d900d0930fa9721d111e217d07a0fcc193f5a689910b8cd5d3f5d2","abstract_canon_sha256":"016bb668ae7163a5606a58208693ef6cd762090aa10277c64d701b35821cbfab"},"schema_version":"1.0"},"canonical_sha256":"c454464b58288d9e636daa87c80ff01ccb7657ed832370301a81bcef57f68626","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:59:54.644718Z","signature_b64":"+klgxtFGci+mmwgGJcrHwrr94Qe3mSiloXsuRvPhHQDh+ib8WfGMKa4wLQwSFszTmrcdzy87Ooa0kKnqRv+3Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c454464b58288d9e636daa87c80ff01ccb7657ed832370301a81bcef57f68626","last_reissued_at":"2026-07-05T08:59:54.644298Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:59:54.644298Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2408.15116","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":"2026-07-05T08:59:54Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"tJm6e9WtUODBNltZ+dGOaPdqlXo63Kowdh+8V/VwibgLP0etrvK3sjLX+TEgGHYQ9BKIuOCJZceowAGjXBltCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-19T02:15:20.918725Z"},"content_sha256":"233335442a62cfc739b1e80a73ec4e37deb27bd030784df1e8d16b43985c445d","schema_version":"1.0","event_id":"sha256:233335442a62cfc739b1e80a73ec4e37deb27bd030784df1e8d16b43985c445d"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:YRKEMS2YFCGZ4Y3NVKD4QD7QDT","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Evaluating Stability of Unreflective Alignment","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"James Lucassen, Mark Henry, Owen Yeung, Philippa Wright","submitted_at":"2024-08-27T14:55:15Z","abstract_excerpt":"Many theoretical obstacles to AI alignment are consequences of reflective stability - the problem of designing alignment mechanisms that the AI would not disable if given the option. However, problems stemming from reflective stability are not obviously present in current LLMs, leading to disagreement over whether they will need to be solved to enable safe delegation of cognitive labor. In this paper, we propose Counterfactual Priority Change (CPC) destabilization as a mechanism by which reflective stability problems may arise in future LLMs. We describe two risk factors for CPC-destabilizatio"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.15116","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/2408.15116/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":"2026-07-05T08:59:54Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"hPf8QIMaHeIT4NtyMm7cPhGQ3BzhvmzzI9pmPTVa2JVOaiRkMudu7y9jGMFgeM7/IRV6Sc6SvP1d9vze3uInBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-19T02:15:20.919231Z"},"content_sha256":"a787be3665889bba78a68a706c6f9f06cff6591e9a78f90a6c8be3009c6b809a","schema_version":"1.0","event_id":"sha256:a787be3665889bba78a68a706c6f9f06cff6591e9a78f90a6c8be3009c6b809a"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/YRKEMS2YFCGZ4Y3NVKD4QD7QDT/bundle.json","state_url":"https://pith.science/pith/YRKEMS2YFCGZ4Y3NVKD4QD7QDT/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/YRKEMS2YFCGZ4Y3NVKD4QD7QDT/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":"2026-08-19T02:15:20Z","links":{"resolver":"https://pith.science/pith/YRKEMS2YFCGZ4Y3NVKD4QD7QDT","bundle":"https://pith.science/pith/YRKEMS2YFCGZ4Y3NVKD4QD7QDT/bundle.json","state":"https://pith.science/pith/YRKEMS2YFCGZ4Y3NVKD4QD7QDT/state.json","well_known_bundle":"https://pith.science/.well-known/pith/YRKEMS2YFCGZ4Y3NVKD4QD7QDT/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:YRKEMS2YFCGZ4Y3NVKD4QD7QDT","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":"016bb668ae7163a5606a58208693ef6cd762090aa10277c64d701b35821cbfab","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-08-27T14:55:15Z","title_canon_sha256":"926e331176d900d0930fa9721d111e217d07a0fcc193f5a689910b8cd5d3f5d2"},"schema_version":"1.0","source":{"id":"2408.15116","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2408.15116","created_at":"2026-07-05T08:59:54Z"},{"alias_kind":"arxiv_version","alias_value":"2408.15116v1","created_at":"2026-07-05T08:59:54Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.15116","created_at":"2026-07-05T08:59:54Z"},{"alias_kind":"pith_short_12","alias_value":"YRKEMS2YFCGZ","created_at":"2026-07-05T08:59:54Z"},{"alias_kind":"pith_short_16","alias_value":"YRKEMS2YFCGZ4Y3N","created_at":"2026-07-05T08:59:54Z"},{"alias_kind":"pith_short_8","alias_value":"YRKEMS2Y","created_at":"2026-07-05T08:59:54Z"}],"graph_snapshots":[{"event_id":"sha256:a787be3665889bba78a68a706c6f9f06cff6591e9a78f90a6c8be3009c6b809a","target":"graph","created_at":"2026-07-05T08:59:54Z","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/2408.15116/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Many theoretical obstacles to AI alignment are consequences of reflective stability - the problem of designing alignment mechanisms that the AI would not disable if given the option. However, problems stemming from reflective stability are not obviously present in current LLMs, leading to disagreement over whether they will need to be solved to enable safe delegation of cognitive labor. In this paper, we propose Counterfactual Priority Change (CPC) destabilization as a mechanism by which reflective stability problems may arise in future LLMs. We describe two risk factors for CPC-destabilizatio","authors_text":"James Lucassen, Mark Henry, Owen Yeung, Philippa Wright","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-08-27T14:55:15Z","title":"Evaluating Stability of Unreflective Alignment"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.15116","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":"sha256:233335442a62cfc739b1e80a73ec4e37deb27bd030784df1e8d16b43985c445d","target":"record","created_at":"2026-07-05T08:59:54Z","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":"016bb668ae7163a5606a58208693ef6cd762090aa10277c64d701b35821cbfab","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-08-27T14:55:15Z","title_canon_sha256":"926e331176d900d0930fa9721d111e217d07a0fcc193f5a689910b8cd5d3f5d2"},"schema_version":"1.0","source":{"id":"2408.15116","kind":"arxiv","version":1}},"canonical_sha256":"c454464b58288d9e636daa87c80ff01ccb7657ed832370301a81bcef57f68626","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"c454464b58288d9e636daa87c80ff01ccb7657ed832370301a81bcef57f68626","first_computed_at":"2026-07-05T08:59:54.644298Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:59:54.644298Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"+klgxtFGci+mmwgGJcrHwrr94Qe3mSiloXsuRvPhHQDh+ib8WfGMKa4wLQwSFszTmrcdzy87Ooa0kKnqRv+3Bw==","signature_status":"signed_v1","signed_at":"2026-07-05T08:59:54.644718Z","signed_message":"canonical_sha256_bytes"},"source_id":"2408.15116","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:233335442a62cfc739b1e80a73ec4e37deb27bd030784df1e8d16b43985c445d","sha256:a787be3665889bba78a68a706c6f9f06cff6591e9a78f90a6c8be3009c6b809a"],"state_sha256":"9386b503dffd9d25d1f6eef5450321ed21827be7e0b7072ab2cf39d0e7f610d4"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"G9OghKuV07zRw+t8uR7XlOSSH/ZG7qtZVBaSuGpeMFM57bIgB/5Z5GUmx1+h1JluXavuehfrXagXPeiEZ/UuDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-19T02:15:20.923288Z","bundle_sha256":"5dd8d64cd7a8259c69643851d34799c2a1f92ab38512dd7d8073df3a756e79b5"}}