{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:NOJQKM3XIM63V653DVOSQHNK45","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":"9f4ea1443e40ca9330407563f52e36f65870cee25f041a4e6ca843f4a6e313c1","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-08-28T13:22:33Z","title_canon_sha256":"381ffc1dd98a7f4fb5ffa48a6e8821f006fa47872c90a732b04928be30044039"},"schema_version":"1.0","source":{"id":"2508.20766","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.20766","created_at":"2026-07-05T12:01:10Z"},{"alias_kind":"arxiv_version","alias_value":"2508.20766v1","created_at":"2026-07-05T12:01:10Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.20766","created_at":"2026-07-05T12:01:10Z"},{"alias_kind":"pith_short_12","alias_value":"NOJQKM3XIM63","created_at":"2026-07-05T12:01:10Z"},{"alias_kind":"pith_short_16","alias_value":"NOJQKM3XIM63V653","created_at":"2026-07-05T12:01:10Z"},{"alias_kind":"pith_short_8","alias_value":"NOJQKM3X","created_at":"2026-07-05T12:01:10Z"}],"graph_snapshots":[{"event_id":"sha256:1b56ccadecacefa69b38b1be9ea210001e1486f9f299b40d3ae3488474086516","target":"graph","created_at":"2026-07-05T12:01:10Z","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/2508.20766/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Safety alignment in Large Language Models (LLMs) often involves mediating internal representations to refuse harmful requests. Recent research has demonstrated that these safety mechanisms can be bypassed by ablating or removing specific representational directions within the model. In this paper, we propose the opposite approach: Rank-One Safety Injection (ROSI), a white-box method that amplifies a model's safety alignment by permanently steering its activations toward the refusal-mediating subspace. ROSI operates as a simple, fine-tuning-free rank-one weight modification applied to all resid","authors_text":"Bernard Ghanem, George Turkiyyah, Harethah Abu Shairah, Hasan Abed Al Kader Hammoud","cross_cats":["cs.AI","cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-08-28T13:22:33Z","title":"Turning the Spell Around: Lightweight Alignment Amplification via Rank-One Safety Injection"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.20766","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":"sha256:50c86705414f13ac3213f63e1bd094ecf5b24005d2ee24e7a22b501d0e1dc9c7","target":"record","created_at":"2026-07-05T12:01:10Z","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":"9f4ea1443e40ca9330407563f52e36f65870cee25f041a4e6ca843f4a6e313c1","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-08-28T13:22:33Z","title_canon_sha256":"381ffc1dd98a7f4fb5ffa48a6e8821f006fa47872c90a732b04928be30044039"},"schema_version":"1.0","source":{"id":"2508.20766","kind":"arxiv","version":1}},"canonical_sha256":"6b93053377433dbafbbb1d5d281daae772dda6382f124bb69c8bda3add2dee7c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"6b93053377433dbafbbb1d5d281daae772dda6382f124bb69c8bda3add2dee7c","first_computed_at":"2026-07-05T12:01:10.484608Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T12:01:10.484608Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"59DLawDrlN/qtZpux2DJA5yl7cA1zTxuR2KJKUyFE8wPUlA8kb2C6Zt1YHCslkxBoiN6W9x+QAD9Icp23GKuBA==","signature_status":"signed_v1","signed_at":"2026-07-05T12:01:10.485167Z","signed_message":"canonical_sha256_bytes"},"source_id":"2508.20766","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:50c86705414f13ac3213f63e1bd094ecf5b24005d2ee24e7a22b501d0e1dc9c7","sha256:1b56ccadecacefa69b38b1be9ea210001e1486f9f299b40d3ae3488474086516"],"state_sha256":"426b20c01170b710fd634711908c80ab9e8757610a7c0ab0c14293ffbd88c1aa"}