{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:YDQYDRTDM3WYTNQJ6O53KR4YFT","short_pith_number":"pith:YDQYDRTD","canonical_record":{"source":{"id":"2606.16349","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CR","submitted_at":"2026-06-15T07:50:00Z","cross_cats_sorted":[],"title_canon_sha256":"cc84ba117be3cfd6e32641505b64d991ce7942d65887edf7d6bfbb31df522f2a","abstract_canon_sha256":"df9a81ffa48bedc397c360e876a44fd10a9605746adc66938fe173f0a0bbedc4"},"schema_version":"1.0"},"canonical_sha256":"c0e181c66366ed89b609f3bbb547982cfea3f21a8934f3ff9de53ebcde25d4e3","source":{"kind":"arxiv","id":"2606.16349","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.16349","created_at":"2026-06-19T16:10:01Z"},{"alias_kind":"arxiv_version","alias_value":"2606.16349v2","created_at":"2026-06-19T16:10:01Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.16349","created_at":"2026-06-19T16:10:01Z"},{"alias_kind":"pith_short_12","alias_value":"YDQYDRTDM3WY","created_at":"2026-06-19T16:10:01Z"},{"alias_kind":"pith_short_16","alias_value":"YDQYDRTDM3WYTNQJ","created_at":"2026-06-19T16:10:01Z"},{"alias_kind":"pith_short_8","alias_value":"YDQYDRTD","created_at":"2026-06-19T16:10:01Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:YDQYDRTDM3WYTNQJ6O53KR4YFT","target":"record","payload":{"canonical_record":{"source":{"id":"2606.16349","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CR","submitted_at":"2026-06-15T07:50:00Z","cross_cats_sorted":[],"title_canon_sha256":"cc84ba117be3cfd6e32641505b64d991ce7942d65887edf7d6bfbb31df522f2a","abstract_canon_sha256":"df9a81ffa48bedc397c360e876a44fd10a9605746adc66938fe173f0a0bbedc4"},"schema_version":"1.0"},"canonical_sha256":"c0e181c66366ed89b609f3bbb547982cfea3f21a8934f3ff9de53ebcde25d4e3","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-19T16:10:01.802077Z","signature_b64":"E53zTEqS7iBprwcsgeaH6EwMNNR7sLiNZ5rrdjog+di763cjuEhfs2kfMhG9nEh59LdBnfP/hOUEsNASAlZdAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c0e181c66366ed89b609f3bbb547982cfea3f21a8934f3ff9de53ebcde25d4e3","last_reissued_at":"2026-06-19T16:10:01.801678Z","signature_status":"signed_v1","first_computed_at":"2026-06-19T16:10:01.801678Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2606.16349","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-19T16:10:01Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"mKSk+HVfWfdL9mvbFLd8NNj41jPO4BrWrplfivDern5ukkxTWW58nNXKObbAdqsTAncKMIkwAtElpUbZ50LsBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-22T23:57:18.167008Z"},"content_sha256":"def3405267c41f9af406098265d9e25a2e4f39c8b8df627f9fed17878c74a426","schema_version":"1.0","event_id":"sha256:def3405267c41f9af406098265d9e25a2e4f39c8b8df627f9fed17878c74a426"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:YDQYDRTDM3WYTNQJ6O53KR4YFT","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"From Refusal Geometry to Safety Geometry: Harmfulness--Refusal Coupling under Dynamic Adversarial Fine-Tuning","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Haihua Shen, Junbin Yang, Meiqi Wu, Shan Li, Wenhao Lan, Xinhua Lai, Yijun Yang","submitted_at":"2026-06-15T07:50:00Z","abstract_excerpt":"Safety alignment requires language models to refuse harmful requests without losing the ability to answer benign ones. Existing robustness evaluations, however, do not reveal whether a model has learned to recognize harmfulness, to activate a refusal policy, or to couple these two processes. We study this question with a dual safety-geometry protocol that measures harmfulness carriers, refusal carriers, and their coupling across aligned instruction-tuned anchors and matched Mistral-7B-v0.1 SFT/R2D2 training trajectories. The aligned anchors validate the protocol: refusal-side interventions reo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.16349","kind":"arxiv","version":2},"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/2606.16349/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":"2026-06-19T16:10:01Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"EzePIesdDCnOU6/JDx5Zfujh+sZWw5K/1JnTR3n7HcLAmXseURLmLDbeYPjEoCXWjFxw+qEmVtVnoEkuMY1YAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-22T23:57:18.167409Z"},"content_sha256":"be3697ef0e9dfbe0adc671b9c7eecf4f1c3ee3fcda34e80338ec91058aeeb02d","schema_version":"1.0","event_id":"sha256:be3697ef0e9dfbe0adc671b9c7eecf4f1c3ee3fcda34e80338ec91058aeeb02d"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/YDQYDRTDM3WYTNQJ6O53KR4YFT/bundle.json","state_url":"https://pith.science/pith/YDQYDRTDM3WYTNQJ6O53KR4YFT/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/YDQYDRTDM3WYTNQJ6O53KR4YFT/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":"2026-08-22T23:57:18Z","links":{"resolver":"https://pith.science/pith/YDQYDRTDM3WYTNQJ6O53KR4YFT","bundle":"https://pith.science/pith/YDQYDRTDM3WYTNQJ6O53KR4YFT/bundle.json","state":"https://pith.science/pith/YDQYDRTDM3WYTNQJ6O53KR4YFT/state.json","well_known_bundle":"https://pith.science/.well-known/pith/YDQYDRTDM3WYTNQJ6O53KR4YFT/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:YDQYDRTDM3WYTNQJ6O53KR4YFT","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":"df9a81ffa48bedc397c360e876a44fd10a9605746adc66938fe173f0a0bbedc4","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CR","submitted_at":"2026-06-15T07:50:00Z","title_canon_sha256":"cc84ba117be3cfd6e32641505b64d991ce7942d65887edf7d6bfbb31df522f2a"},"schema_version":"1.0","source":{"id":"2606.16349","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.16349","created_at":"2026-06-19T16:10:01Z"},{"alias_kind":"arxiv_version","alias_value":"2606.16349v2","created_at":"2026-06-19T16:10:01Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.16349","created_at":"2026-06-19T16:10:01Z"},{"alias_kind":"pith_short_12","alias_value":"YDQYDRTDM3WY","created_at":"2026-06-19T16:10:01Z"},{"alias_kind":"pith_short_16","alias_value":"YDQYDRTDM3WYTNQJ","created_at":"2026-06-19T16:10:01Z"},{"alias_kind":"pith_short_8","alias_value":"YDQYDRTD","created_at":"2026-06-19T16:10:01Z"}],"graph_snapshots":[{"event_id":"sha256:be3697ef0e9dfbe0adc671b9c7eecf4f1c3ee3fcda34e80338ec91058aeeb02d","target":"graph","created_at":"2026-06-19T16:10:01Z","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/2606.16349/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Safety alignment requires language models to refuse harmful requests without losing the ability to answer benign ones. Existing robustness evaluations, however, do not reveal whether a model has learned to recognize harmfulness, to activate a refusal policy, or to couple these two processes. We study this question with a dual safety-geometry protocol that measures harmfulness carriers, refusal carriers, and their coupling across aligned instruction-tuned anchors and matched Mistral-7B-v0.1 SFT/R2D2 training trajectories. The aligned anchors validate the protocol: refusal-side interventions reo","authors_text":"Haihua Shen, Junbin Yang, Meiqi Wu, Shan Li, Wenhao Lan, Xinhua Lai, Yijun Yang","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CR","submitted_at":"2026-06-15T07:50:00Z","title":"From Refusal Geometry to Safety Geometry: Harmfulness--Refusal Coupling under Dynamic Adversarial Fine-Tuning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.16349","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:def3405267c41f9af406098265d9e25a2e4f39c8b8df627f9fed17878c74a426","target":"record","created_at":"2026-06-19T16:10:01Z","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":"df9a81ffa48bedc397c360e876a44fd10a9605746adc66938fe173f0a0bbedc4","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CR","submitted_at":"2026-06-15T07:50:00Z","title_canon_sha256":"cc84ba117be3cfd6e32641505b64d991ce7942d65887edf7d6bfbb31df522f2a"},"schema_version":"1.0","source":{"id":"2606.16349","kind":"arxiv","version":2}},"canonical_sha256":"c0e181c66366ed89b609f3bbb547982cfea3f21a8934f3ff9de53ebcde25d4e3","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"c0e181c66366ed89b609f3bbb547982cfea3f21a8934f3ff9de53ebcde25d4e3","first_computed_at":"2026-06-19T16:10:01.801678Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-19T16:10:01.801678Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"E53zTEqS7iBprwcsgeaH6EwMNNR7sLiNZ5rrdjog+di763cjuEhfs2kfMhG9nEh59LdBnfP/hOUEsNASAlZdAg==","signature_status":"signed_v1","signed_at":"2026-06-19T16:10:01.802077Z","signed_message":"canonical_sha256_bytes"},"source_id":"2606.16349","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:def3405267c41f9af406098265d9e25a2e4f39c8b8df627f9fed17878c74a426","sha256:be3697ef0e9dfbe0adc671b9c7eecf4f1c3ee3fcda34e80338ec91058aeeb02d"],"state_sha256":"6f0e3f9b51b9772820b06f42adee55bb5e94337c3753803bf32cf7c0c3b94c72"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/0HsUJU05K35/9XMkv2ss9t4B3KMCuqObQBDyCscd7N2mAtMG9jVclVwWhXjBJZglhSjjDhvBo4NRa6O+Bq3Cg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-22T23:57:18.170811Z","bundle_sha256":"73d3ac16ad8d21b76c7d453eecfc05ebe13e092edd1b2312daf2000952ce4a1f"}}