{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:3FSGNBUVSRZXJ3ZHYDIKIZBDZ4","short_pith_number":"pith:3FSGNBUV","schema_version":"1.0","canonical_sha256":"d964668695947374ef27c0d0a46423cf1ba893b9494ebae6bb8d4b888fdbdf93","source":{"kind":"arxiv","id":"2402.10207","version":6},"attestation_state":"computed","paper":{"title":"Rewards-in-Context: Multi-objective Alignment of Foundation Models with Dynamic Preference Adjustment","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Dong Yu, Feng Luo, Han Zhong, Jianshu Chen, Rui Yang, Shuang Qiu, Xiaoman Pan","submitted_at":"2024-02-15T18:58:31Z","abstract_excerpt":"We consider the problem of multi-objective alignment of foundation models with human preferences, which is a critical step towards helpful and harmless AI systems. However, it is generally costly and unstable to fine-tune large foundation models using reinforcement learning (RL), and the multi-dimensionality, heterogeneity, and conflicting nature of human preferences further complicate the alignment process. In this paper, we introduce Rewards-in-Context (RiC), which conditions the response of a foundation model on multiple rewards in its prompt context and applies supervised fine-tuning for a"},"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":"2402.10207","kind":"arxiv","version":6},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-02-15T18:58:31Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"08128ac739bf0198d5670b6d1fc86ab00974d65e2e3785a7acc205bc6468436b","abstract_canon_sha256":"8c1614fabe669f8587361d9b69e498b1dd0f3cee4967cba536e439ef7af34da5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:21:02.693742Z","signature_b64":"LI52r76sMyQpabHtXNQOo3V6ph7k+Y+NnjNbHMhLgmdbM1511v8R+Jn7PEbiqswYNyxLn5FNG08YscLhy+U7AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d964668695947374ef27c0d0a46423cf1ba893b9494ebae6bb8d4b888fdbdf93","last_reissued_at":"2026-07-05T09:21:02.693222Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:21:02.693222Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Rewards-in-Context: Multi-objective Alignment of Foundation Models with Dynamic Preference Adjustment","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Dong Yu, Feng Luo, Han Zhong, Jianshu Chen, Rui Yang, Shuang Qiu, Xiaoman Pan","submitted_at":"2024-02-15T18:58:31Z","abstract_excerpt":"We consider the problem of multi-objective alignment of foundation models with human preferences, which is a critical step towards helpful and harmless AI systems. However, it is generally costly and unstable to fine-tune large foundation models using reinforcement learning (RL), and the multi-dimensionality, heterogeneity, and conflicting nature of human preferences further complicate the alignment process. In this paper, we introduce Rewards-in-Context (RiC), which conditions the response of a foundation model on multiple rewards in its prompt context and applies supervised fine-tuning for a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.10207","kind":"arxiv","version":6},"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/2402.10207/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":"2402.10207","created_at":"2026-07-05T09:21:02.693288+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.10207v6","created_at":"2026-07-05T09:21:02.693288+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.10207","created_at":"2026-07-05T09:21:02.693288+00:00"},{"alias_kind":"pith_short_12","alias_value":"3FSGNBUVSRZX","created_at":"2026-07-05T09:21:02.693288+00:00"},{"alias_kind":"pith_short_16","alias_value":"3FSGNBUVSRZXJ3ZH","created_at":"2026-07-05T09:21:02.693288+00:00"},{"alias_kind":"pith_short_8","alias_value":"3FSGNBUV","created_at":"2026-07-05T09:21:02.693288+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.07988","citing_title":"PAFO: Pareto Fairness Optimization for Personalized Reward Modeling","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20619","citing_title":"SURF: Steering the Scalarization Weight to Uniformly Traverse the Pareto Front","ref_index":104,"is_internal_anchor":false},{"citing_arxiv_id":"2603.18113","citing_title":"VC-Soup: Value-Consistency Guided Multi-Value Alignment for Large Language Models","ref_index":46,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11679","citing_title":"Explaining and Breaking the Safety-Helpfulness Ceiling via Preference Dimensional Expansion","ref_index":65,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11679","citing_title":"Explaining and Breaking the Safety-Helpfulness Ceiling via Preference Dimensional Expansion","ref_index":65,"is_internal_anchor":false},{"citing_arxiv_id":"2604.27699","citing_title":"Bridging Values and Behavior: A Hierarchical Framework for Proactive Embodied Agents","ref_index":50,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3FSGNBUVSRZXJ3ZHYDIKIZBDZ4","json":"https://pith.science/pith/3FSGNBUVSRZXJ3ZHYDIKIZBDZ4.json","graph_json":"https://pith.science/api/pith-number/3FSGNBUVSRZXJ3ZHYDIKIZBDZ4/graph.json","events_json":"https://pith.science/api/pith-number/3FSGNBUVSRZXJ3ZHYDIKIZBDZ4/events.json","paper":"https://pith.science/paper/3FSGNBUV"},"agent_actions":{"view_html":"https://pith.science/pith/3FSGNBUVSRZXJ3ZHYDIKIZBDZ4","download_json":"https://pith.science/pith/3FSGNBUVSRZXJ3ZHYDIKIZBDZ4.json","view_paper":"https://pith.science/paper/3FSGNBUV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.10207&json=true","fetch_graph":"https://pith.science/api/pith-number/3FSGNBUVSRZXJ3ZHYDIKIZBDZ4/graph.json","fetch_events":"https://pith.science/api/pith-number/3FSGNBUVSRZXJ3ZHYDIKIZBDZ4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3FSGNBUVSRZXJ3ZHYDIKIZBDZ4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3FSGNBUVSRZXJ3ZHYDIKIZBDZ4/action/storage_attestation","attest_author":"https://pith.science/pith/3FSGNBUVSRZXJ3ZHYDIKIZBDZ4/action/author_attestation","sign_citation":"https://pith.science/pith/3FSGNBUVSRZXJ3ZHYDIKIZBDZ4/action/citation_signature","submit_replication":"https://pith.science/pith/3FSGNBUVSRZXJ3ZHYDIKIZBDZ4/action/replication_record"}},"created_at":"2026-07-05T09:21:02.693288+00:00","updated_at":"2026-07-05T09:21:02.693288+00:00"}