{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:CJTHLIWNLGUPZYOHKLXD4UUUX4","short_pith_number":"pith:CJTHLIWN","schema_version":"1.0","canonical_sha256":"126675a2cd59a8fce1c752ee3e5294bf3be82a3e7e1949f23bf81ecdbc9886be","source":{"kind":"arxiv","id":"2409.13733","version":1},"attestation_state":"computed","paper":{"title":"RNR: Teaching Large Language Models to Follow Roles and Rules","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.HC"],"primary_cat":"cs.CL","authors_text":"Alexander Bukharin, Bing Yin, Chao Zhang, Haoming Jiang, Jianshu Chen, Jingbo Shang, Kuan Wang, Qingyu Yin, Shiyang Li, Tuo Zhao, Xian Li, Zhengyang Wang","submitted_at":"2024-09-10T06:07:32Z","abstract_excerpt":"Instruction fine-tuning (IFT) elicits instruction following capabilities and steers the behavior of large language models (LLMs) via supervised learning. However, existing models trained on open-source IFT datasets only have the ability to follow instructions from users, and often fail to follow complex role and rules specified by developers, a.k.a. system prompts. The ability to follow these roles and rules is essential for deployment, as it ensures that the model safely interacts with users within developer defined guidelines. To improve such role and rule following ability, we propose \\mode"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2409.13733","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-09-10T06:07:32Z","cross_cats_sorted":["cs.AI","cs.HC"],"title_canon_sha256":"b5ed750e45900ac89fc740bae0c4185cac6398a703591c803f9c0b20c37eec51","abstract_canon_sha256":"73fc25611803dbf3f68634fef400dbe6778d26f4ab716bd4e975de489a2000c2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:10:32.875230Z","signature_b64":"96SWZ3LsKRv6E6tpvqUqqxf4qfG6qxx10dqzNvApYhmRqA7xaW5VIvOSBKOYtMEeSs2zNdpKjR9ZC8URcFMoCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"126675a2cd59a8fce1c752ee3e5294bf3be82a3e7e1949f23bf81ecdbc9886be","last_reissued_at":"2026-07-05T09:10:32.874821Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:10:32.874821Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"RNR: Teaching Large Language Models to Follow Roles and Rules","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.HC"],"primary_cat":"cs.CL","authors_text":"Alexander Bukharin, Bing Yin, Chao Zhang, Haoming Jiang, Jianshu Chen, Jingbo Shang, Kuan Wang, Qingyu Yin, Shiyang Li, Tuo Zhao, Xian Li, Zhengyang Wang","submitted_at":"2024-09-10T06:07:32Z","abstract_excerpt":"Instruction fine-tuning (IFT) elicits instruction following capabilities and steers the behavior of large language models (LLMs) via supervised learning. However, existing models trained on open-source IFT datasets only have the ability to follow instructions from users, and often fail to follow complex role and rules specified by developers, a.k.a. system prompts. The ability to follow these roles and rules is essential for deployment, as it ensures that the model safely interacts with users within developer defined guidelines. To improve such role and rule following ability, we propose \\mode"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.13733","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/2409.13733/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2409.13733","created_at":"2026-07-05T09:10:32.874877+00:00"},{"alias_kind":"arxiv_version","alias_value":"2409.13733v1","created_at":"2026-07-05T09:10:32.874877+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.13733","created_at":"2026-07-05T09:10:32.874877+00:00"},{"alias_kind":"pith_short_12","alias_value":"CJTHLIWNLGUP","created_at":"2026-07-05T09:10:32.874877+00:00"},{"alias_kind":"pith_short_16","alias_value":"CJTHLIWNLGUPZYOH","created_at":"2026-07-05T09:10:32.874877+00:00"},{"alias_kind":"pith_short_8","alias_value":"CJTHLIWN","created_at":"2026-07-05T09:10:32.874877+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.26955","citing_title":"Policy-Governed LLM Routing with Intent Matching for Instrument Laboratories","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06196","citing_title":"The Granularity Axis: A Micro-to-Macro Latent Direction for Social Roles in Language Models","ref_index":72,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CJTHLIWNLGUPZYOHKLXD4UUUX4","json":"https://pith.science/pith/CJTHLIWNLGUPZYOHKLXD4UUUX4.json","graph_json":"https://pith.science/api/pith-number/CJTHLIWNLGUPZYOHKLXD4UUUX4/graph.json","events_json":"https://pith.science/api/pith-number/CJTHLIWNLGUPZYOHKLXD4UUUX4/events.json","paper":"https://pith.science/paper/CJTHLIWN"},"agent_actions":{"view_html":"https://pith.science/pith/CJTHLIWNLGUPZYOHKLXD4UUUX4","download_json":"https://pith.science/pith/CJTHLIWNLGUPZYOHKLXD4UUUX4.json","view_paper":"https://pith.science/paper/CJTHLIWN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2409.13733&json=true","fetch_graph":"https://pith.science/api/pith-number/CJTHLIWNLGUPZYOHKLXD4UUUX4/graph.json","fetch_events":"https://pith.science/api/pith-number/CJTHLIWNLGUPZYOHKLXD4UUUX4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CJTHLIWNLGUPZYOHKLXD4UUUX4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CJTHLIWNLGUPZYOHKLXD4UUUX4/action/storage_attestation","attest_author":"https://pith.science/pith/CJTHLIWNLGUPZYOHKLXD4UUUX4/action/author_attestation","sign_citation":"https://pith.science/pith/CJTHLIWNLGUPZYOHKLXD4UUUX4/action/citation_signature","submit_replication":"https://pith.science/pith/CJTHLIWNLGUPZYOHKLXD4UUUX4/action/replication_record"}},"created_at":"2026-07-05T09:10:32.874877+00:00","updated_at":"2026-07-05T09:10:32.874877+00:00"}