{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:L6L3YHU5QIQUM5OL4GJ4T6OMAV","short_pith_number":"pith:L6L3YHU5","canonical_record":{"source":{"id":"2406.08124","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-12T12:06:32Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"23c2fcdde08e5fc96adf243d9acc9c079aa8a0f555cde2932d6bb80a3c5fa824","abstract_canon_sha256":"de33d0c138cb4b19743ca124ceb86e7a39ab397cd769e7757920ef179c7fd473"},"schema_version":"1.0"},"canonical_sha256":"5f97bc1e9d82214675cbe193c9f9cc05598e1ee89d429563671cb27c5cc7329f","source":{"kind":"arxiv","id":"2406.08124","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.08124","created_at":"2026-07-05T09:50:48Z"},{"alias_kind":"arxiv_version","alias_value":"2406.08124v2","created_at":"2026-07-05T09:50:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.08124","created_at":"2026-07-05T09:50:48Z"},{"alias_kind":"pith_short_12","alias_value":"L6L3YHU5QIQU","created_at":"2026-07-05T09:50:48Z"},{"alias_kind":"pith_short_16","alias_value":"L6L3YHU5QIQUM5OL","created_at":"2026-07-05T09:50:48Z"},{"alias_kind":"pith_short_8","alias_value":"L6L3YHU5","created_at":"2026-07-05T09:50:48Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:L6L3YHU5QIQUM5OL4GJ4T6OMAV","target":"record","payload":{"canonical_record":{"source":{"id":"2406.08124","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-12T12:06:32Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"23c2fcdde08e5fc96adf243d9acc9c079aa8a0f555cde2932d6bb80a3c5fa824","abstract_canon_sha256":"de33d0c138cb4b19743ca124ceb86e7a39ab397cd769e7757920ef179c7fd473"},"schema_version":"1.0"},"canonical_sha256":"5f97bc1e9d82214675cbe193c9f9cc05598e1ee89d429563671cb27c5cc7329f","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:50:48.414287Z","signature_b64":"SjuBOKW+JRCkHZatk3CdGM0JS7ngsDa5x5lpqBdUpBDPuMfn1+EnW/zIEB1WxLltJDJOOsBociEzTe5aStJmBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5f97bc1e9d82214675cbe193c9f9cc05598e1ee89d429563671cb27c5cc7329f","last_reissued_at":"2026-07-05T09:50:48.413736Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:50:48.413736Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2406.08124","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":"2026-07-05T09:50:48Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Afci8BU7Kl+/KUallYzaRZLe5nac3pHfkwkS6LkWA72h7dDSFHOxt2bzdrTYbc6oN4Vfmp/nH+wTSFyMIEklBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-19T02:13:47.689181Z"},"content_sha256":"b4c3c13efd42b69a00e77e1b6a886ea74efeac1205a1a3e6843f625bb88a6223","schema_version":"1.0","event_id":"sha256:b4c3c13efd42b69a00e77e1b6a886ea74efeac1205a1a3e6843f625bb88a6223"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:L6L3YHU5QIQUM5OL4GJ4T6OMAV","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Legend: Leveraging Representation Engineering to Annotate Safety Margin for Preference Datasets","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Bowen Qin, Chen Huang, Duanyu Feng, Wenqiang Lei, Youcheng Huang, Zheng Zhang","submitted_at":"2024-06-12T12:06:32Z","abstract_excerpt":"The success of the reward model in distinguishing between responses with subtle safety differences depends critically on the high-quality preference dataset, which should capture the fine-grained nuances of harmful and harmless responses. This motivates the need to develop a dataset involving preference margins, which accurately quantify how harmless one response is compared to another. In this paper, we take the first step to propose an effective and cost-efficient framework to promote the margin-enhanced preference dataset development. Our framework, Legend, Leverages representation engineer"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.08124","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/2406.08124/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":"2026-07-05T09:50:48Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"853nIZ/fRemtfZGICq7zZqdyZui5EoNFUe8DkKTffYfhGjbs5OpFbWwYc0r9XoBZSZRoYjUwJikXyvCrYAzTDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-19T02:13:47.690263Z"},"content_sha256":"02f6228a4471b3c8d7b521264b7f9a58872514374ac3dc56cf4220ce83a5ab15","schema_version":"1.0","event_id":"sha256:02f6228a4471b3c8d7b521264b7f9a58872514374ac3dc56cf4220ce83a5ab15"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/L6L3YHU5QIQUM5OL4GJ4T6OMAV/bundle.json","state_url":"https://pith.science/pith/L6L3YHU5QIQUM5OL4GJ4T6OMAV/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/L6L3YHU5QIQUM5OL4GJ4T6OMAV/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":"2026-08-19T02:13:47Z","links":{"resolver":"https://pith.science/pith/L6L3YHU5QIQUM5OL4GJ4T6OMAV","bundle":"https://pith.science/pith/L6L3YHU5QIQUM5OL4GJ4T6OMAV/bundle.json","state":"https://pith.science/pith/L6L3YHU5QIQUM5OL4GJ4T6OMAV/state.json","well_known_bundle":"https://pith.science/.well-known/pith/L6L3YHU5QIQUM5OL4GJ4T6OMAV/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:L6L3YHU5QIQUM5OL4GJ4T6OMAV","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":"de33d0c138cb4b19743ca124ceb86e7a39ab397cd769e7757920ef179c7fd473","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-12T12:06:32Z","title_canon_sha256":"23c2fcdde08e5fc96adf243d9acc9c079aa8a0f555cde2932d6bb80a3c5fa824"},"schema_version":"1.0","source":{"id":"2406.08124","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.08124","created_at":"2026-07-05T09:50:48Z"},{"alias_kind":"arxiv_version","alias_value":"2406.08124v2","created_at":"2026-07-05T09:50:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.08124","created_at":"2026-07-05T09:50:48Z"},{"alias_kind":"pith_short_12","alias_value":"L6L3YHU5QIQU","created_at":"2026-07-05T09:50:48Z"},{"alias_kind":"pith_short_16","alias_value":"L6L3YHU5QIQUM5OL","created_at":"2026-07-05T09:50:48Z"},{"alias_kind":"pith_short_8","alias_value":"L6L3YHU5","created_at":"2026-07-05T09:50:48Z"}],"graph_snapshots":[{"event_id":"sha256:02f6228a4471b3c8d7b521264b7f9a58872514374ac3dc56cf4220ce83a5ab15","target":"graph","created_at":"2026-07-05T09:50:48Z","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/2406.08124/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The success of the reward model in distinguishing between responses with subtle safety differences depends critically on the high-quality preference dataset, which should capture the fine-grained nuances of harmful and harmless responses. This motivates the need to develop a dataset involving preference margins, which accurately quantify how harmless one response is compared to another. In this paper, we take the first step to propose an effective and cost-efficient framework to promote the margin-enhanced preference dataset development. Our framework, Legend, Leverages representation engineer","authors_text":"Bowen Qin, Chen Huang, Duanyu Feng, Wenqiang Lei, Youcheng Huang, Zheng Zhang","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-12T12:06:32Z","title":"Legend: Leveraging Representation Engineering to Annotate Safety Margin for Preference Datasets"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.08124","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:b4c3c13efd42b69a00e77e1b6a886ea74efeac1205a1a3e6843f625bb88a6223","target":"record","created_at":"2026-07-05T09:50:48Z","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":"de33d0c138cb4b19743ca124ceb86e7a39ab397cd769e7757920ef179c7fd473","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-12T12:06:32Z","title_canon_sha256":"23c2fcdde08e5fc96adf243d9acc9c079aa8a0f555cde2932d6bb80a3c5fa824"},"schema_version":"1.0","source":{"id":"2406.08124","kind":"arxiv","version":2}},"canonical_sha256":"5f97bc1e9d82214675cbe193c9f9cc05598e1ee89d429563671cb27c5cc7329f","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"5f97bc1e9d82214675cbe193c9f9cc05598e1ee89d429563671cb27c5cc7329f","first_computed_at":"2026-07-05T09:50:48.413736Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:50:48.413736Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"SjuBOKW+JRCkHZatk3CdGM0JS7ngsDa5x5lpqBdUpBDPuMfn1+EnW/zIEB1WxLltJDJOOsBociEzTe5aStJmBA==","signature_status":"signed_v1","signed_at":"2026-07-05T09:50:48.414287Z","signed_message":"canonical_sha256_bytes"},"source_id":"2406.08124","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b4c3c13efd42b69a00e77e1b6a886ea74efeac1205a1a3e6843f625bb88a6223","sha256:02f6228a4471b3c8d7b521264b7f9a58872514374ac3dc56cf4220ce83a5ab15"],"state_sha256":"3abade7abf206cf6e644c06daf332253876ac533c7ead2f1e9e8bc7b2d11ab1f"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"+6UFSg4zc3w+M3UZ0ip4WGRWziQzoCCTn03bQTes3TYe/efsKFzXlixqfaUBB8PWmtQnIJ/JGF6mLS/hcxYdBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-19T02:13:47.696854Z","bundle_sha256":"0f96704217a1176daf6c3d49a4aa4d1f1bad0b938248387b69adbcd071209d6e"}}