{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:A2TIPJNKLIHLJAJL6J3ESMV2DL","short_pith_number":"pith:A2TIPJNK","schema_version":"1.0","canonical_sha256":"06a687a5aa5a0eb4812bf2764932ba1af09f4380853f3a0aba089edc9ca8553b","source":{"kind":"arxiv","id":"2312.15907","version":1},"attestation_state":"computed","paper":{"title":"Align on the Fly: Adapting Chatbot Behavior to Established Norms","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chunpu Xu, Ethan Chern, Ge Zhang, Jie Fu, Jing Li, Pengfei Liu, Ruibo Liu, Steffi Chern, Zekun Wang","submitted_at":"2023-12-26T06:51:09Z","abstract_excerpt":"In this paper, we aim to align large language models with the ever-changing, complex, and diverse human values (e.g., social norms) across time and locations. This presents a challenge to existing alignment techniques, such as supervised fine-tuning, which internalize values within model parameters. To overcome this, we propose an On-the-fly Preference Optimization (OPO) method, which is a real-time alignment that works in a streaming way. It employs an external memory to store established rules for alignment, which can constrain LLMs' behaviors without further training, allowing for convenien"},"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":"2312.15907","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-12-26T06:51:09Z","cross_cats_sorted":[],"title_canon_sha256":"75df263f405bebfbf51031acf6b5237000a6b591072a210c81e5bcd91309e2ff","abstract_canon_sha256":"a339d866d4d42dcc1903f13445133652b365572ad93f45deaf8394172c4bc56a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:28:03.335300Z","signature_b64":"6JFFayW8smzPY8YITSj2gjbNQ1bq2tmMI74MVHZB3S/q9pOQIPLq60czFnJLlgCxAT65kYPk7WtFmRR+xWQSCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"06a687a5aa5a0eb4812bf2764932ba1af09f4380853f3a0aba089edc9ca8553b","last_reissued_at":"2026-07-05T07:28:03.334840Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:28:03.334840Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Align on the Fly: Adapting Chatbot Behavior to Established Norms","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chunpu Xu, Ethan Chern, Ge Zhang, Jie Fu, Jing Li, Pengfei Liu, Ruibo Liu, Steffi Chern, Zekun Wang","submitted_at":"2023-12-26T06:51:09Z","abstract_excerpt":"In this paper, we aim to align large language models with the ever-changing, complex, and diverse human values (e.g., social norms) across time and locations. This presents a challenge to existing alignment techniques, such as supervised fine-tuning, which internalize values within model parameters. To overcome this, we propose an On-the-fly Preference Optimization (OPO) method, which is a real-time alignment that works in a streaming way. It employs an external memory to store established rules for alignment, which can constrain LLMs' behaviors without further training, allowing for convenien"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2312.15907","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/2312.15907/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":"2312.15907","created_at":"2026-07-05T07:28:03.334889+00:00"},{"alias_kind":"arxiv_version","alias_value":"2312.15907v1","created_at":"2026-07-05T07:28:03.334889+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2312.15907","created_at":"2026-07-05T07:28:03.334889+00:00"},{"alias_kind":"pith_short_12","alias_value":"A2TIPJNKLIHL","created_at":"2026-07-05T07:28:03.334889+00:00"},{"alias_kind":"pith_short_16","alias_value":"A2TIPJNKLIHLJAJL","created_at":"2026-07-05T07:28:03.334889+00:00"},{"alias_kind":"pith_short_8","alias_value":"A2TIPJNK","created_at":"2026-07-05T07:28:03.334889+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.09016","citing_title":"A Survey on Training-free Alignment of Large Language Models","ref_index":93,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/A2TIPJNKLIHLJAJL6J3ESMV2DL","json":"https://pith.science/pith/A2TIPJNKLIHLJAJL6J3ESMV2DL.json","graph_json":"https://pith.science/api/pith-number/A2TIPJNKLIHLJAJL6J3ESMV2DL/graph.json","events_json":"https://pith.science/api/pith-number/A2TIPJNKLIHLJAJL6J3ESMV2DL/events.json","paper":"https://pith.science/paper/A2TIPJNK"},"agent_actions":{"view_html":"https://pith.science/pith/A2TIPJNKLIHLJAJL6J3ESMV2DL","download_json":"https://pith.science/pith/A2TIPJNKLIHLJAJL6J3ESMV2DL.json","view_paper":"https://pith.science/paper/A2TIPJNK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2312.15907&json=true","fetch_graph":"https://pith.science/api/pith-number/A2TIPJNKLIHLJAJL6J3ESMV2DL/graph.json","fetch_events":"https://pith.science/api/pith-number/A2TIPJNKLIHLJAJL6J3ESMV2DL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/A2TIPJNKLIHLJAJL6J3ESMV2DL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/A2TIPJNKLIHLJAJL6J3ESMV2DL/action/storage_attestation","attest_author":"https://pith.science/pith/A2TIPJNKLIHLJAJL6J3ESMV2DL/action/author_attestation","sign_citation":"https://pith.science/pith/A2TIPJNKLIHLJAJL6J3ESMV2DL/action/citation_signature","submit_replication":"https://pith.science/pith/A2TIPJNKLIHLJAJL6J3ESMV2DL/action/replication_record"}},"created_at":"2026-07-05T07:28:03.334889+00:00","updated_at":"2026-07-05T07:28:03.334889+00:00"}