{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:WSDADBLBKD7NPS34OY7U7OGAE2","short_pith_number":"pith:WSDADBLB","canonical_record":{"source":{"id":"2608.05578","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2026-08-06T03:57:38Z","cross_cats_sorted":[],"title_canon_sha256":"56427d06dd9a94058145ab2b66a0b2d717affeb8f0072becafe86dcd9f6ae6df","abstract_canon_sha256":"dea0eef4ceff681314520f7447bcf81af331c8d91be36ebd9b7a6e23dc53af81"},"schema_version":"1.0"},"canonical_sha256":"b48601856150fed7cb7c763f4fb8c026b2ace327dfcee0bd2f4e4d233a8fc59c","source":{"kind":"arxiv","id":"2608.05578","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2608.05578","created_at":"2026-08-07T00:49:56Z"},{"alias_kind":"arxiv_version","alias_value":"2608.05578v1","created_at":"2026-08-07T00:49:56Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.05578","created_at":"2026-08-07T00:49:56Z"},{"alias_kind":"pith_short_12","alias_value":"WSDADBLBKD7N","created_at":"2026-08-07T00:49:56Z"},{"alias_kind":"pith_short_16","alias_value":"WSDADBLBKD7NPS34","created_at":"2026-08-07T00:49:56Z"},{"alias_kind":"pith_short_8","alias_value":"WSDADBLB","created_at":"2026-08-07T00:49:56Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:WSDADBLBKD7NPS34OY7U7OGAE2","target":"record","payload":{"canonical_record":{"source":{"id":"2608.05578","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2026-08-06T03:57:38Z","cross_cats_sorted":[],"title_canon_sha256":"56427d06dd9a94058145ab2b66a0b2d717affeb8f0072becafe86dcd9f6ae6df","abstract_canon_sha256":"dea0eef4ceff681314520f7447bcf81af331c8d91be36ebd9b7a6e23dc53af81"},"schema_version":"1.0"},"canonical_sha256":"b48601856150fed7cb7c763f4fb8c026b2ace327dfcee0bd2f4e4d233a8fc59c","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-07T00:49:56.170952Z","signature_b64":"9hfSUZCqL6ttaUfCNHef3dO3XJwOaUp0zcxpMXqlKC13Y6TtB/TutgDJQ2rmmZ6CMbh3c22wShrga3iDJuQHDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b48601856150fed7cb7c763f4fb8c026b2ace327dfcee0bd2f4e4d233a8fc59c","last_reissued_at":"2026-08-07T00:49:56.169536Z","signature_status":"signed_v1","first_computed_at":"2026-08-07T00:49:56.169536Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2608.05578","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":"2026-08-07T00:49:56Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"PQZ/mgA65F9qO6v+mkSqpTRe1lyP71eMsEVoFtiw77ge/UT0vZEo6ht6ZCrKmNI+2jzrgyLXPbMsNud+rc3XBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T10:51:33.151012Z"},"content_sha256":"3ada0c4de56e6e1a5c27ae2000d0349e28b147d90bd7e7218efb3b15d3f9e1e1","schema_version":"1.0","event_id":"sha256:3ada0c4de56e6e1a5c27ae2000d0349e28b147d90bd7e7218efb3b15d3f9e1e1"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:WSDADBLBKD7NPS34OY7U7OGAE2","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Detecting Safety Training Modification in Language Models via Activation Analysis","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Glen Messenger","submitted_at":"2026-08-06T03:57:38Z","abstract_excerpt":"We introduce AMS (Activation-based Model Scanner), a tool that detects modifications to safety training in language models by measuring the geometric structure of safety-relevant concepts in activation space. Safety training creates measurable separation between harmful and benign content classes; certain safety modifications collapse or rotate this structure, while others leave it intact. We validate AMS across 14 model configurations spanning 4 architecture families (Llama, Gemma, Qwen, Mistral) and four safety-modification categories (instruction-tuned, base, abliterated, uncensored fine-tu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.05578","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/2608.05578/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":"2026-08-07T00:49:56Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"lKXNamKhLdiwDn1CRljY5zA0xRvPDz4i3Iw/can34B+iDLxlut4XHoYwIvwaiFjCE2pCnKmys6+fzlW5TRJEAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T10:51:33.151514Z"},"content_sha256":"07bc19885720d35b5b31a7d3595d950fd912718d7c5040c297085ca1049a543b","schema_version":"1.0","event_id":"sha256:07bc19885720d35b5b31a7d3595d950fd912718d7c5040c297085ca1049a543b"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/WSDADBLBKD7NPS34OY7U7OGAE2/bundle.json","state_url":"https://pith.science/pith/WSDADBLBKD7NPS34OY7U7OGAE2/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/WSDADBLBKD7NPS34OY7U7OGAE2/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":"2026-08-09T10:51:33Z","links":{"resolver":"https://pith.science/pith/WSDADBLBKD7NPS34OY7U7OGAE2","bundle":"https://pith.science/pith/WSDADBLBKD7NPS34OY7U7OGAE2/bundle.json","state":"https://pith.science/pith/WSDADBLBKD7NPS34OY7U7OGAE2/state.json","well_known_bundle":"https://pith.science/.well-known/pith/WSDADBLBKD7NPS34OY7U7OGAE2/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:WSDADBLBKD7NPS34OY7U7OGAE2","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":"dea0eef4ceff681314520f7447bcf81af331c8d91be36ebd9b7a6e23dc53af81","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2026-08-06T03:57:38Z","title_canon_sha256":"56427d06dd9a94058145ab2b66a0b2d717affeb8f0072becafe86dcd9f6ae6df"},"schema_version":"1.0","source":{"id":"2608.05578","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2608.05578","created_at":"2026-08-07T00:49:56Z"},{"alias_kind":"arxiv_version","alias_value":"2608.05578v1","created_at":"2026-08-07T00:49:56Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.05578","created_at":"2026-08-07T00:49:56Z"},{"alias_kind":"pith_short_12","alias_value":"WSDADBLBKD7N","created_at":"2026-08-07T00:49:56Z"},{"alias_kind":"pith_short_16","alias_value":"WSDADBLBKD7NPS34","created_at":"2026-08-07T00:49:56Z"},{"alias_kind":"pith_short_8","alias_value":"WSDADBLB","created_at":"2026-08-07T00:49:56Z"}],"graph_snapshots":[{"event_id":"sha256:07bc19885720d35b5b31a7d3595d950fd912718d7c5040c297085ca1049a543b","target":"graph","created_at":"2026-08-07T00:49:56Z","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/2608.05578/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We introduce AMS (Activation-based Model Scanner), a tool that detects modifications to safety training in language models by measuring the geometric structure of safety-relevant concepts in activation space. Safety training creates measurable separation between harmful and benign content classes; certain safety modifications collapse or rotate this structure, while others leave it intact. We validate AMS across 14 model configurations spanning 4 architecture families (Llama, Gemma, Qwen, Mistral) and four safety-modification categories (instruction-tuned, base, abliterated, uncensored fine-tu","authors_text":"Glen Messenger","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2026-08-06T03:57:38Z","title":"Detecting Safety Training Modification in Language Models via Activation Analysis"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.05578","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":"sha256:3ada0c4de56e6e1a5c27ae2000d0349e28b147d90bd7e7218efb3b15d3f9e1e1","target":"record","created_at":"2026-08-07T00:49:56Z","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":"dea0eef4ceff681314520f7447bcf81af331c8d91be36ebd9b7a6e23dc53af81","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2026-08-06T03:57:38Z","title_canon_sha256":"56427d06dd9a94058145ab2b66a0b2d717affeb8f0072becafe86dcd9f6ae6df"},"schema_version":"1.0","source":{"id":"2608.05578","kind":"arxiv","version":1}},"canonical_sha256":"b48601856150fed7cb7c763f4fb8c026b2ace327dfcee0bd2f4e4d233a8fc59c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"b48601856150fed7cb7c763f4fb8c026b2ace327dfcee0bd2f4e4d233a8fc59c","first_computed_at":"2026-08-07T00:49:56.169536Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-08-07T00:49:56.169536Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"9hfSUZCqL6ttaUfCNHef3dO3XJwOaUp0zcxpMXqlKC13Y6TtB/TutgDJQ2rmmZ6CMbh3c22wShrga3iDJuQHDQ==","signature_status":"signed_v1","signed_at":"2026-08-07T00:49:56.170952Z","signed_message":"canonical_sha256_bytes"},"source_id":"2608.05578","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3ada0c4de56e6e1a5c27ae2000d0349e28b147d90bd7e7218efb3b15d3f9e1e1","sha256:07bc19885720d35b5b31a7d3595d950fd912718d7c5040c297085ca1049a543b"],"state_sha256":"8c2118b7b308c498d39e2dec7c3cff7e3d320a48283a52723d6c2e7b906ab509"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Zuxe60ECaGpWsfmb6yXgigF5pDx3gTPFk2oqNlEEYrbdh6/HbtFMZxftMmaO71qe0Nu35i3AvD+lwM1gJdqEAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T10:51:33.156019Z","bundle_sha256":"37e99ecfd6974a9854c45a5a45bbedd5e310dd289983734424fa4c65920260d1"}}