{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:AK53QFNGCKVFZJMPRLYMP5PVUM","short_pith_number":"pith:AK53QFNG","schema_version":"1.0","canonical_sha256":"02bbb815a612aa5ca58f8af0c7f5f5a32f518c5b728cfdf61a2a711ae175ac53","source":{"kind":"arxiv","id":"2606.07705","version":1},"attestation_state":"computed","paper":{"title":"SAW: Stage-Aware Dynamic Weighting for Multi-Objective Reinforcement Learning in Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Baolong Bi, Bolin Wan, Huaming Liao, Jiafeng Guo, Juan Chen, Shenghua Liu, Siqian Tong, Xueqi Cheng, Yuchen He, Yuyao Ge","submitted_at":"2026-06-05T10:00:19Z","abstract_excerpt":"Although multi-objective reinforcement learning (MORL) is central to aligning large language models with complex human preferences, the prevailing practice of static weighted summation overlooks a more fundamental phenomenon: reward learning is markedly asynchronous across objectives. Well-learned dimensions quickly produce homogeneous, low-variance signals whose residual noise contaminates the aggregated reward (in GRPO) or occupies a fixed share of the advantage budget (in GDPO), interfering with the scarce yet high-value signals carried by under-learned dimensions. To address this asynchron"},"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":"2606.07705","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-06-05T10:00:19Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"d61a14e73fcca1dab14159bffbfc4fa0e20f3073d56eb82b7d1a548e7e8c65a9","abstract_canon_sha256":"5eae22ef2f6c1defbd29f76d73606ee9b8079cb6d42dcdc55c7d47bbca5c8fc7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-09T01:04:49.798545Z","signature_b64":"5baiTzm7CnqfahcENpL+2dMwvtpFeCT9ofMet2RLgPB3qCFTFIxhL0L7ZZRm7PKejW4Dj0f7jASAwUtkpEH5BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"02bbb815a612aa5ca58f8af0c7f5f5a32f518c5b728cfdf61a2a711ae175ac53","last_reissued_at":"2026-06-09T01:04:49.798047Z","signature_status":"signed_v1","first_computed_at":"2026-06-09T01:04:49.798047Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SAW: Stage-Aware Dynamic Weighting for Multi-Objective Reinforcement Learning in Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Baolong Bi, Bolin Wan, Huaming Liao, Jiafeng Guo, Juan Chen, Shenghua Liu, Siqian Tong, Xueqi Cheng, Yuchen He, Yuyao Ge","submitted_at":"2026-06-05T10:00:19Z","abstract_excerpt":"Although multi-objective reinforcement learning (MORL) is central to aligning large language models with complex human preferences, the prevailing practice of static weighted summation overlooks a more fundamental phenomenon: reward learning is markedly asynchronous across objectives. Well-learned dimensions quickly produce homogeneous, low-variance signals whose residual noise contaminates the aggregated reward (in GRPO) or occupies a fixed share of the advantage budget (in GDPO), interfering with the scarce yet high-value signals carried by under-learned dimensions. To address this asynchron"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.07705","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/2606.07705/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":"2606.07705","created_at":"2026-06-09T01:04:49.798097+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.07705v1","created_at":"2026-06-09T01:04:49.798097+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.07705","created_at":"2026-06-09T01:04:49.798097+00:00"},{"alias_kind":"pith_short_12","alias_value":"AK53QFNGCKVF","created_at":"2026-06-09T01:04:49.798097+00:00"},{"alias_kind":"pith_short_16","alias_value":"AK53QFNGCKVFZJMP","created_at":"2026-06-09T01:04:49.798097+00:00"},{"alias_kind":"pith_short_8","alias_value":"AK53QFNG","created_at":"2026-06-09T01:04:49.798097+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AK53QFNGCKVFZJMPRLYMP5PVUM","json":"https://pith.science/pith/AK53QFNGCKVFZJMPRLYMP5PVUM.json","graph_json":"https://pith.science/api/pith-number/AK53QFNGCKVFZJMPRLYMP5PVUM/graph.json","events_json":"https://pith.science/api/pith-number/AK53QFNGCKVFZJMPRLYMP5PVUM/events.json","paper":"https://pith.science/paper/AK53QFNG"},"agent_actions":{"view_html":"https://pith.science/pith/AK53QFNGCKVFZJMPRLYMP5PVUM","download_json":"https://pith.science/pith/AK53QFNGCKVFZJMPRLYMP5PVUM.json","view_paper":"https://pith.science/paper/AK53QFNG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.07705&json=true","fetch_graph":"https://pith.science/api/pith-number/AK53QFNGCKVFZJMPRLYMP5PVUM/graph.json","fetch_events":"https://pith.science/api/pith-number/AK53QFNGCKVFZJMPRLYMP5PVUM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AK53QFNGCKVFZJMPRLYMP5PVUM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AK53QFNGCKVFZJMPRLYMP5PVUM/action/storage_attestation","attest_author":"https://pith.science/pith/AK53QFNGCKVFZJMPRLYMP5PVUM/action/author_attestation","sign_citation":"https://pith.science/pith/AK53QFNGCKVFZJMPRLYMP5PVUM/action/citation_signature","submit_replication":"https://pith.science/pith/AK53QFNGCKVFZJMPRLYMP5PVUM/action/replication_record"}},"created_at":"2026-06-09T01:04:49.798097+00:00","updated_at":"2026-06-09T01:04:49.798097+00:00"}