{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:HKKPGHWLJ7YRTQ4W6SZS7D4D6P","short_pith_number":"pith:HKKPGHWL","canonical_record":{"source":{"id":"2509.06982","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-09-01T04:50:02Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"19dc1c850741d7fc1471c732416dc66d79d8d921c87942fe2e7235d47c6d069a","abstract_canon_sha256":"ee44d47901acbef5e3db26001b6b93b351aeb5742a73679b28df64a77dc9b777"},"schema_version":"1.0"},"canonical_sha256":"3a94f31ecb4ff119c396f4b32f8f83f3e704916ab7280119e650fb970ba4a8b8","source":{"kind":"arxiv","id":"2509.06982","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2509.06982","created_at":"2026-07-05T12:07:13Z"},{"alias_kind":"arxiv_version","alias_value":"2509.06982v1","created_at":"2026-07-05T12:07:13Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.06982","created_at":"2026-07-05T12:07:13Z"},{"alias_kind":"pith_short_12","alias_value":"HKKPGHWLJ7YR","created_at":"2026-07-05T12:07:13Z"},{"alias_kind":"pith_short_16","alias_value":"HKKPGHWLJ7YRTQ4W","created_at":"2026-07-05T12:07:13Z"},{"alias_kind":"pith_short_8","alias_value":"HKKPGHWL","created_at":"2026-07-05T12:07:13Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:HKKPGHWLJ7YRTQ4W6SZS7D4D6P","target":"record","payload":{"canonical_record":{"source":{"id":"2509.06982","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-09-01T04:50:02Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"19dc1c850741d7fc1471c732416dc66d79d8d921c87942fe2e7235d47c6d069a","abstract_canon_sha256":"ee44d47901acbef5e3db26001b6b93b351aeb5742a73679b28df64a77dc9b777"},"schema_version":"1.0"},"canonical_sha256":"3a94f31ecb4ff119c396f4b32f8f83f3e704916ab7280119e650fb970ba4a8b8","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:07:13.451802Z","signature_b64":"fdxPUtBtp+ax2iai3pwZ4ci/aGwQeYGny4pGDsay508ga133s2qujqItTj3ZvpKJro0QQrNvkKLfD+JUOLIxDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3a94f31ecb4ff119c396f4b32f8f83f3e704916ab7280119e650fb970ba4a8b8","last_reissued_at":"2026-07-05T12:07:13.451392Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:07:13.451392Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2509.06982","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":"2026-07-05T12:07:13Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"uSyIDe2srUtwRJORdM9ALWHtyb5YMPfpYIk4H7WpZq/sbKZMylzE7Cye8D4gRaVhrpKCn+eBbJTLIywK2wH6Ag==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-17T17:38:00.279777Z"},"content_sha256":"de082211665c62c8ce172710dd34857e6fa855082206343cda160eba63666d79","schema_version":"1.0","event_id":"sha256:de082211665c62c8ce172710dd34857e6fa855082206343cda160eba63666d79"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:HKKPGHWLJ7YRTQ4W6SZS7D4D6P","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"CARE: Decoding Time Safety Alignment via Rollback and Introspection Intervention","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Chenhan Yuan, Fei Huang, Junyang Lin, Tsung-Yi Ho, Xiaomeng Hu","submitted_at":"2025-09-01T04:50:02Z","abstract_excerpt":"As large language models (LLMs) are increasingly deployed in real-world applications, ensuring the safety of their outputs during decoding has become a critical challenge. However, existing decoding-time interventions, such as Contrastive Decoding, often force a severe trade-off between safety and response quality. In this work, we propose CARE, a novel framework for decoding-time safety alignment that integrates three key components: (1) a guard model for real-time safety monitoring, enabling detection of potentially unsafe content; (2) a rollback mechanism with a token buffer to correct unsa"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.06982","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/2509.06982/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":"2026-07-05T12:07:13Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Gvy38Kqe68eb9ititgq1jurUlwENX9Z2xsiOHWmpR56g1/tQRl6xJEEH1MvYPK5kPcvYZuk0ruAvgEMgHDLMBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-17T17:38:00.280898Z"},"content_sha256":"cd8a60a44c6106603bf6826958a2c6c74c63e5992b84f56528bf965de328f1b3","schema_version":"1.0","event_id":"sha256:cd8a60a44c6106603bf6826958a2c6c74c63e5992b84f56528bf965de328f1b3"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/HKKPGHWLJ7YRTQ4W6SZS7D4D6P/bundle.json","state_url":"https://pith.science/pith/HKKPGHWLJ7YRTQ4W6SZS7D4D6P/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/HKKPGHWLJ7YRTQ4W6SZS7D4D6P/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":"2026-08-17T17:38:00Z","links":{"resolver":"https://pith.science/pith/HKKPGHWLJ7YRTQ4W6SZS7D4D6P","bundle":"https://pith.science/pith/HKKPGHWLJ7YRTQ4W6SZS7D4D6P/bundle.json","state":"https://pith.science/pith/HKKPGHWLJ7YRTQ4W6SZS7D4D6P/state.json","well_known_bundle":"https://pith.science/.well-known/pith/HKKPGHWLJ7YRTQ4W6SZS7D4D6P/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:HKKPGHWLJ7YRTQ4W6SZS7D4D6P","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":"ee44d47901acbef5e3db26001b6b93b351aeb5742a73679b28df64a77dc9b777","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-09-01T04:50:02Z","title_canon_sha256":"19dc1c850741d7fc1471c732416dc66d79d8d921c87942fe2e7235d47c6d069a"},"schema_version":"1.0","source":{"id":"2509.06982","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2509.06982","created_at":"2026-07-05T12:07:13Z"},{"alias_kind":"arxiv_version","alias_value":"2509.06982v1","created_at":"2026-07-05T12:07:13Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.06982","created_at":"2026-07-05T12:07:13Z"},{"alias_kind":"pith_short_12","alias_value":"HKKPGHWLJ7YR","created_at":"2026-07-05T12:07:13Z"},{"alias_kind":"pith_short_16","alias_value":"HKKPGHWLJ7YRTQ4W","created_at":"2026-07-05T12:07:13Z"},{"alias_kind":"pith_short_8","alias_value":"HKKPGHWL","created_at":"2026-07-05T12:07:13Z"}],"graph_snapshots":[{"event_id":"sha256:cd8a60a44c6106603bf6826958a2c6c74c63e5992b84f56528bf965de328f1b3","target":"graph","created_at":"2026-07-05T12:07:13Z","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/2509.06982/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"As large language models (LLMs) are increasingly deployed in real-world applications, ensuring the safety of their outputs during decoding has become a critical challenge. However, existing decoding-time interventions, such as Contrastive Decoding, often force a severe trade-off between safety and response quality. In this work, we propose CARE, a novel framework for decoding-time safety alignment that integrates three key components: (1) a guard model for real-time safety monitoring, enabling detection of potentially unsafe content; (2) a rollback mechanism with a token buffer to correct unsa","authors_text":"Chenhan Yuan, Fei Huang, Junyang Lin, Tsung-Yi Ho, Xiaomeng Hu","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-09-01T04:50:02Z","title":"CARE: Decoding Time Safety Alignment via Rollback and Introspection Intervention"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.06982","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:de082211665c62c8ce172710dd34857e6fa855082206343cda160eba63666d79","target":"record","created_at":"2026-07-05T12:07:13Z","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":"ee44d47901acbef5e3db26001b6b93b351aeb5742a73679b28df64a77dc9b777","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-09-01T04:50:02Z","title_canon_sha256":"19dc1c850741d7fc1471c732416dc66d79d8d921c87942fe2e7235d47c6d069a"},"schema_version":"1.0","source":{"id":"2509.06982","kind":"arxiv","version":1}},"canonical_sha256":"3a94f31ecb4ff119c396f4b32f8f83f3e704916ab7280119e650fb970ba4a8b8","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"3a94f31ecb4ff119c396f4b32f8f83f3e704916ab7280119e650fb970ba4a8b8","first_computed_at":"2026-07-05T12:07:13.451392Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T12:07:13.451392Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"fdxPUtBtp+ax2iai3pwZ4ci/aGwQeYGny4pGDsay508ga133s2qujqItTj3ZvpKJro0QQrNvkKLfD+JUOLIxDw==","signature_status":"signed_v1","signed_at":"2026-07-05T12:07:13.451802Z","signed_message":"canonical_sha256_bytes"},"source_id":"2509.06982","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:de082211665c62c8ce172710dd34857e6fa855082206343cda160eba63666d79","sha256:cd8a60a44c6106603bf6826958a2c6c74c63e5992b84f56528bf965de328f1b3"],"state_sha256":"6bbdb65b98e6be57684c51204d0247aae04362a22587c0e3941605983c7251fb"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7zUV+IsWAsDUB0YzEs5YTz8tCIgtbM8guHwfEyYzrzET5TBuXEAmb7t151w2spbrSXHCROgfg7e8kDRGb4ZFCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-17T17:38:00.289893Z","bundle_sha256":"37e6718b20e9eedea229439d815d86b47130ca71bc0b7cb7e20f4eb10ec591bc"}}