{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:RGCX4YG6TIPMIEEBXOLKTMRO5K","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":"a30eeb0146adc8ec63861c7fd2a5ef2ba6b1dbef9386d8dbd7708de408fd50e0","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-12T18:38:05Z","title_canon_sha256":"47a85d26e73ac3faf8d84e7091771c2d3b8394f5a693d7e3b9d3ae06dfd0b301"},"schema_version":"1.0","source":{"id":"2605.12645","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.12645","created_at":"2026-05-18T03:09:59Z"},{"alias_kind":"arxiv_version","alias_value":"2605.12645v1","created_at":"2026-05-18T03:09:59Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.12645","created_at":"2026-05-18T03:09:59Z"},{"alias_kind":"pith_short_12","alias_value":"RGCX4YG6TIPM","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"RGCX4YG6TIPMIEEB","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"RGCX4YG6","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:9f5e4237768f16b3b4139393b1ca7bad2dd7f838f9fa294ecc6da0e44081190a","target":"graph","created_at":"2026-05-18T03:09:59Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Through experiments on the LaMP-QA benchmark across six models, IAP consistently outperforms all baselines, achieving an average macro-score gain of around 7.5% over the strongest competitor."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That a tag-based schema combined with a personalized reward function can reliably infer implicit intent from single-turn questions and optimize generation paths to produce better-aligned answers without multi-turn context or user profiles."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"IAP uses RL to train LLMs to explicitly infer and apply implicit user intent in single-turn personalized QA, achieving ~7.5% average macro-score gains over baselines on LaMP-QA."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Reinforcement learning trains LLMs to infer implicit user intent from single-turn questions and generate better-aligned personalized answers."}],"snapshot_sha256":"dfc034d118c321f692f23d3343f39fac88d2f93d01aab786fa0beab25b22160d"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"Effective personalized question answering (PQA) in language models requires grounding responses in the user's underlying intent, where intent refers to the implicit ``why'' behind a query beyond its explicit wording. However, existing approaches to intent-aware personalization rely on multi-turn conversational context or rich user profiles, and do not explicitly model user intent during the reasoning process. This limits their effectiveness in single-turn settings, where the user's latent goal must be inferred from minimal input and integrated into the thinking and reasoning process. To bridge","authors_text":"Benjamin Charles Germain Lee, Jevin West, Maryam Amirizaniani, Nicholas Weber","cross_cats":["cs.AI"],"headline":"Reinforcement learning trains LLMs to infer implicit user intent from single-turn questions and generate better-aligned personalized answers.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-12T18:38:05Z","title":"Training LLMs with Reinforcement Learning for Intent-Aware Personalized Question Answering"},"references":{"count":91,"internal_anchors":9,"resolved_work":91,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Second Conference on Language Modeling , year=","work_id":"ffcdbe1d-35f0-4dc4-a785-0b468dbbc059","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Yixin Ye and Zhen Huang and Yang Xiao and Ethan Chern and Shijie Xia and Pengfei Liu , booktitle=. 2025 , url=","work_id":"ad06628e-c7a4-4e8f-9893-bb30e26e388d","year":2025},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Search-R1: Training","work_id":"ff26bde5-9a7f-4f44-93bd-690d4d4352ba","year":2025},{"cited_arxiv_id":"","doi":"10.18653/v1/2025.findings-emnlp.84","is_internal_anchor":false,"ref_index":4,"title":"Legal Mathematical Reasoning with LLM s: Procedural Alignment through Two-Stage Reinforcement Learning","work_id":"dea66ccf-46fa-4f16-b72e-15198d93a238","year":2025},{"cited_arxiv_id":"","doi":"10.18653/v1/2025.findings-emnlp.95","is_internal_anchor":false,"ref_index":5,"title":"Beyond Guilt: Legal Judgment Prediction with Trichotomous Reasoning","work_id":"ed49b423-0235-4879-9953-3463dda20901","year":2025}],"snapshot_sha256":"28bb72d7dae523175e4b1f8838f492128c21e20d99b7f191f51f26b576ad8af6"},"source":{"id":"2605.12645","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-14T20:54:25.568151Z","id":"19f2dc69-3520-4bfa-98fe-6c6c88e005c6","model_set":{"reader":"grok-4.3"},"one_line_summary":"IAP uses RL to train LLMs to explicitly infer and apply implicit user intent in single-turn personalized QA, achieving ~7.5% average macro-score gains over baselines on LaMP-QA.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Reinforcement learning trains LLMs to infer implicit user intent from single-turn questions and generate better-aligned personalized answers.","strongest_claim":"Through experiments on the LaMP-QA benchmark across six models, IAP consistently outperforms all baselines, achieving an average macro-score gain of around 7.5% over the strongest competitor.","weakest_assumption":"That a tag-based schema combined with a personalized reward function can reliably infer implicit intent from single-turn questions and optimize generation paths to produce better-aligned answers without multi-turn context or user profiles."}},"verdict_id":"19f2dc69-3520-4bfa-98fe-6c6c88e005c6"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:e6d5b50003269d71e5ae9608565fe7fc7fca80b0ea661fcf28b18c82a681922c","target":"record","created_at":"2026-05-18T03:09:59Z","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":"a30eeb0146adc8ec63861c7fd2a5ef2ba6b1dbef9386d8dbd7708de408fd50e0","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-12T18:38:05Z","title_canon_sha256":"47a85d26e73ac3faf8d84e7091771c2d3b8394f5a693d7e3b9d3ae06dfd0b301"},"schema_version":"1.0","source":{"id":"2605.12645","kind":"arxiv","version":1}},"canonical_sha256":"89857e60de9a1ec41081bb96a9b22eeaa39ecb81e7ae0e10f885c80ea4faaf8b","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"89857e60de9a1ec41081bb96a9b22eeaa39ecb81e7ae0e10f885c80ea4faaf8b","first_computed_at":"2026-05-18T03:09:59.860930Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T03:09:59.860930Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"DOWksHX9c4I+ptEq+h6dXHbTZddTR4ZBMSHMSFfB9g9H1y9UWOXzP9uK/I02Bl5SSx3Qk9o37vqun48juVE8Bg==","signature_status":"signed_v1","signed_at":"2026-05-18T03:09:59.861482Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.12645","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:e6d5b50003269d71e5ae9608565fe7fc7fca80b0ea661fcf28b18c82a681922c","sha256:9f5e4237768f16b3b4139393b1ca7bad2dd7f838f9fa294ecc6da0e44081190a"],"state_sha256":"4c07fc4ef92172d085074eb147a9a2fe974c34c96f36606b18463a1434ade4dd"}