{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:IIXE4FOU7DUZRGIYUIWLBKCTVZ","short_pith_number":"pith:IIXE4FOU","canonical_record":{"source":{"id":"2410.10014","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-13T21:24:25Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"33dcbad41928f423b10af9640fccc5cec7e2d302cca26bc02efa52bc1312886d","abstract_canon_sha256":"3229a73bf65c72beba95dfdc361481446a9a7d69b3226777d612f84e9ff2194e"},"schema_version":"1.0"},"canonical_sha256":"422e4e15d4f8e9989918a22cb0a853ae58b48f6ae7b4cde75673c76f388640df","source":{"kind":"arxiv","id":"2410.10014","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.10014","created_at":"2026-07-05T09:20:09Z"},{"alias_kind":"arxiv_version","alias_value":"2410.10014v1","created_at":"2026-07-05T09:20:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.10014","created_at":"2026-07-05T09:20:09Z"},{"alias_kind":"pith_short_12","alias_value":"IIXE4FOU7DUZ","created_at":"2026-07-05T09:20:09Z"},{"alias_kind":"pith_short_16","alias_value":"IIXE4FOU7DUZRGIY","created_at":"2026-07-05T09:20:09Z"},{"alias_kind":"pith_short_8","alias_value":"IIXE4FOU","created_at":"2026-07-05T09:20:09Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:IIXE4FOU7DUZRGIYUIWLBKCTVZ","target":"record","payload":{"canonical_record":{"source":{"id":"2410.10014","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-13T21:24:25Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"33dcbad41928f423b10af9640fccc5cec7e2d302cca26bc02efa52bc1312886d","abstract_canon_sha256":"3229a73bf65c72beba95dfdc361481446a9a7d69b3226777d612f84e9ff2194e"},"schema_version":"1.0"},"canonical_sha256":"422e4e15d4f8e9989918a22cb0a853ae58b48f6ae7b4cde75673c76f388640df","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:20:09.473481Z","signature_b64":"9BUZ++/DfEUkzd9Lx80rwN4OuQQYSTz/9hPSa4O9ee2hQwSp9KB1ujwoQMO1uDmyQUU0HlmOp7gvUxw0iGZwBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"422e4e15d4f8e9989918a22cb0a853ae58b48f6ae7b4cde75673c76f388640df","last_reissued_at":"2026-07-05T09:20:09.473041Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:20:09.473041Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2410.10014","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":"2026-07-05T09:20:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"0uq8TTIqUoOY7uDF5K1DuxmCwXbfvXPKFssoTobXJvS1cvbNcLnG9rKOCykwd0W4QwRjfmWSt5lRm017WJ3rCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T21:17:05.469406Z"},"content_sha256":"b344a30cfda064512acd5d7422715ec43cdf746369293473aec2b08ba77e6a89","schema_version":"1.0","event_id":"sha256:b344a30cfda064512acd5d7422715ec43cdf746369293473aec2b08ba77e6a89"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:IIXE4FOU7DUZRGIYUIWLBKCTVZ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Safety-Aware Fine-Tuning of Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Hyeong Kyu Choi, Xuefeng Du, Yixuan Li","submitted_at":"2024-10-13T21:24:25Z","abstract_excerpt":"Fine-tuning Large Language Models (LLMs) has emerged as a common practice for tailoring models to individual needs and preferences. The choice of datasets for fine-tuning can be diverse, introducing safety concerns regarding the potential inclusion of harmful data samples. Manually filtering or avoiding such samples, however, can be labor-intensive and subjective. To address these difficulties, we propose a novel Safety-Aware Fine-Tuning (SAFT) framework designed to automatically detect and remove potentially harmful data, by leveraging a scoring function that exploits the subspace information"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.10014","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/2410.10014/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":"2026-07-05T09:20:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"MeHMbhV/iUHrzT4luYaTDQs16EYUULjyeY7mOqnVhqyJgiM8XSUwgSOY6slXUPwvsors4Eu4dKBg5EO8y7hvDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T21:17:05.469920Z"},"content_sha256":"ae59705b8af8563f1ca40b3d85620292e4731804250cd2c4c5f3820081171e2c","schema_version":"1.0","event_id":"sha256:ae59705b8af8563f1ca40b3d85620292e4731804250cd2c4c5f3820081171e2c"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/IIXE4FOU7DUZRGIYUIWLBKCTVZ/bundle.json","state_url":"https://pith.science/pith/IIXE4FOU7DUZRGIYUIWLBKCTVZ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/IIXE4FOU7DUZRGIYUIWLBKCTVZ/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":"2026-08-07T21:17:05Z","links":{"resolver":"https://pith.science/pith/IIXE4FOU7DUZRGIYUIWLBKCTVZ","bundle":"https://pith.science/pith/IIXE4FOU7DUZRGIYUIWLBKCTVZ/bundle.json","state":"https://pith.science/pith/IIXE4FOU7DUZRGIYUIWLBKCTVZ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/IIXE4FOU7DUZRGIYUIWLBKCTVZ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:IIXE4FOU7DUZRGIYUIWLBKCTVZ","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":"3229a73bf65c72beba95dfdc361481446a9a7d69b3226777d612f84e9ff2194e","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-13T21:24:25Z","title_canon_sha256":"33dcbad41928f423b10af9640fccc5cec7e2d302cca26bc02efa52bc1312886d"},"schema_version":"1.0","source":{"id":"2410.10014","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.10014","created_at":"2026-07-05T09:20:09Z"},{"alias_kind":"arxiv_version","alias_value":"2410.10014v1","created_at":"2026-07-05T09:20:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.10014","created_at":"2026-07-05T09:20:09Z"},{"alias_kind":"pith_short_12","alias_value":"IIXE4FOU7DUZ","created_at":"2026-07-05T09:20:09Z"},{"alias_kind":"pith_short_16","alias_value":"IIXE4FOU7DUZRGIY","created_at":"2026-07-05T09:20:09Z"},{"alias_kind":"pith_short_8","alias_value":"IIXE4FOU","created_at":"2026-07-05T09:20:09Z"}],"graph_snapshots":[{"event_id":"sha256:ae59705b8af8563f1ca40b3d85620292e4731804250cd2c4c5f3820081171e2c","target":"graph","created_at":"2026-07-05T09:20:09Z","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/2410.10014/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Fine-tuning Large Language Models (LLMs) has emerged as a common practice for tailoring models to individual needs and preferences. The choice of datasets for fine-tuning can be diverse, introducing safety concerns regarding the potential inclusion of harmful data samples. Manually filtering or avoiding such samples, however, can be labor-intensive and subjective. To address these difficulties, we propose a novel Safety-Aware Fine-Tuning (SAFT) framework designed to automatically detect and remove potentially harmful data, by leveraging a scoring function that exploits the subspace information","authors_text":"Hyeong Kyu Choi, Xuefeng Du, Yixuan Li","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-13T21:24:25Z","title":"Safety-Aware Fine-Tuning of Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.10014","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:b344a30cfda064512acd5d7422715ec43cdf746369293473aec2b08ba77e6a89","target":"record","created_at":"2026-07-05T09:20:09Z","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":"3229a73bf65c72beba95dfdc361481446a9a7d69b3226777d612f84e9ff2194e","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-13T21:24:25Z","title_canon_sha256":"33dcbad41928f423b10af9640fccc5cec7e2d302cca26bc02efa52bc1312886d"},"schema_version":"1.0","source":{"id":"2410.10014","kind":"arxiv","version":1}},"canonical_sha256":"422e4e15d4f8e9989918a22cb0a853ae58b48f6ae7b4cde75673c76f388640df","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"422e4e15d4f8e9989918a22cb0a853ae58b48f6ae7b4cde75673c76f388640df","first_computed_at":"2026-07-05T09:20:09.473041Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:20:09.473041Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"9BUZ++/DfEUkzd9Lx80rwN4OuQQYSTz/9hPSa4O9ee2hQwSp9KB1ujwoQMO1uDmyQUU0HlmOp7gvUxw0iGZwBQ==","signature_status":"signed_v1","signed_at":"2026-07-05T09:20:09.473481Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.10014","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b344a30cfda064512acd5d7422715ec43cdf746369293473aec2b08ba77e6a89","sha256:ae59705b8af8563f1ca40b3d85620292e4731804250cd2c4c5f3820081171e2c"],"state_sha256":"ff4e91f6c972f3da275fb1a04097057d2201eda077feaccef3896dd1bbc65c12"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"LpCTjPxfD9eREMsqOTgZB0E/hVSmAMS/FC8rPXoIs3vdqlXCSzOSfWqs2VrsZxH1URBXcklBjv8vwXl6IRqNCA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-07T21:17:05.475147Z","bundle_sha256":"5bc6696c5493c5d692e161d7b6abd490e3f06158713784074e4e1134fc484842"}}