{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:624IDQGHL3PAV7MTZNXJ32FOWH","short_pith_number":"pith:624IDQGH","schema_version":"1.0","canonical_sha256":"f6b881c0c75ede0afd93cb6e9de8aeb1e1b482d268908b7ffb83b9cd9abf6932","source":{"kind":"arxiv","id":"2606.11634","version":1},"attestation_state":"computed","paper":{"title":"Architecture-Aware Reinforcement Learning Makes Sliding-Window Attention Competitive in Math Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Jianfei Gao, Kai Chen, Kai Liu, Peijie Dong, Qipeng Guo, Shaoting Zhang, Xiaowen Chu, Xinchen Xie","submitted_at":"2026-06-10T03:56:03Z","abstract_excerpt":"The rapid progress of reasoning and agentic large language models (LLMs) has increased the demand for long-context inference, but self-attention (SA) scales quadratically with context length.\n  To address this, we study SWARR (Sliding-Window Attention with Reinforced Adaptation for Math Reasoning), a practical recipe for adapting SWA models to mathematical reasoning. SWARR has two stages: (1) efficient conversion from a pretrained SA model to SWA with supervised fine-tuning (SFT), which avoids pretraining a new base model, and (2) policy adaptation with reinforcement learning (RL).\n  We find t"},"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":"2606.11634","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-06-10T03:56:03Z","cross_cats_sorted":[],"title_canon_sha256":"e3b47809be5cb0a3730b649983d57a6f862ec2ef429a6150c047fb77a04f4608","abstract_canon_sha256":"7ef64d51a53a810f2a603e2b5e7a9fae6dc386bc7c8feabd162af7efda606e10"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-11T01:10:00.087289Z","signature_b64":"jCn3AGBX0UdqcFdg5mr28ud1A7nz64jJsc89MuO5hI8rzwem/Jr/7DwcNsq4uA4N514e56WC8W8tL39p5UoECQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f6b881c0c75ede0afd93cb6e9de8aeb1e1b482d268908b7ffb83b9cd9abf6932","last_reissued_at":"2026-06-11T01:10:00.086821Z","signature_status":"signed_v1","first_computed_at":"2026-06-11T01:10:00.086821Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Architecture-Aware Reinforcement Learning Makes Sliding-Window Attention Competitive in Math Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Jianfei Gao, Kai Chen, Kai Liu, Peijie Dong, Qipeng Guo, Shaoting Zhang, Xiaowen Chu, Xinchen Xie","submitted_at":"2026-06-10T03:56:03Z","abstract_excerpt":"The rapid progress of reasoning and agentic large language models (LLMs) has increased the demand for long-context inference, but self-attention (SA) scales quadratically with context length.\n  To address this, we study SWARR (Sliding-Window Attention with Reinforced Adaptation for Math Reasoning), a practical recipe for adapting SWA models to mathematical reasoning. SWARR has two stages: (1) efficient conversion from a pretrained SA model to SWA with supervised fine-tuning (SFT), which avoids pretraining a new base model, and (2) policy adaptation with reinforcement learning (RL).\n  We find t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.11634","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/2606.11634/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":"2606.11634","created_at":"2026-06-11T01:10:00.086895+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.11634v1","created_at":"2026-06-11T01:10:00.086895+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.11634","created_at":"2026-06-11T01:10:00.086895+00:00"},{"alias_kind":"pith_short_12","alias_value":"624IDQGHL3PA","created_at":"2026-06-11T01:10:00.086895+00:00"},{"alias_kind":"pith_short_16","alias_value":"624IDQGHL3PAV7MT","created_at":"2026-06-11T01:10:00.086895+00:00"},{"alias_kind":"pith_short_8","alias_value":"624IDQGH","created_at":"2026-06-11T01:10:00.086895+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/624IDQGHL3PAV7MTZNXJ32FOWH","json":"https://pith.science/pith/624IDQGHL3PAV7MTZNXJ32FOWH.json","graph_json":"https://pith.science/api/pith-number/624IDQGHL3PAV7MTZNXJ32FOWH/graph.json","events_json":"https://pith.science/api/pith-number/624IDQGHL3PAV7MTZNXJ32FOWH/events.json","paper":"https://pith.science/paper/624IDQGH"},"agent_actions":{"view_html":"https://pith.science/pith/624IDQGHL3PAV7MTZNXJ32FOWH","download_json":"https://pith.science/pith/624IDQGHL3PAV7MTZNXJ32FOWH.json","view_paper":"https://pith.science/paper/624IDQGH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.11634&json=true","fetch_graph":"https://pith.science/api/pith-number/624IDQGHL3PAV7MTZNXJ32FOWH/graph.json","fetch_events":"https://pith.science/api/pith-number/624IDQGHL3PAV7MTZNXJ32FOWH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/624IDQGHL3PAV7MTZNXJ32FOWH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/624IDQGHL3PAV7MTZNXJ32FOWH/action/storage_attestation","attest_author":"https://pith.science/pith/624IDQGHL3PAV7MTZNXJ32FOWH/action/author_attestation","sign_citation":"https://pith.science/pith/624IDQGHL3PAV7MTZNXJ32FOWH/action/citation_signature","submit_replication":"https://pith.science/pith/624IDQGHL3PAV7MTZNXJ32FOWH/action/replication_record"}},"created_at":"2026-06-11T01:10:00.086895+00:00","updated_at":"2026-06-11T01:10:00.086895+00:00"}