{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:2EWTQHJYYEZCWN2COMHX5WHDC4","short_pith_number":"pith:2EWTQHJY","schema_version":"1.0","canonical_sha256":"d12d381d38c1322b3742730f7ed8e3173d6588eefe31a6ae4f61e28e706add79","source":{"kind":"arxiv","id":"2606.27705","version":1},"attestation_state":"computed","paper":{"title":"Mitigating Position Bias in Transformers via Layer-Specific Positional Embedding Scaling","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Changze Lv, Muling Wu, Qi Qian, Shizheng Li, Tianlong Li, Tianyuan Shi, Xiaoqing Zheng, Xuanjing Huang, Yiran Ding, Yixin Wu, Zhenghua Wang, Zhibo Xu","submitted_at":"2026-06-26T04:07:41Z","abstract_excerpt":"Large Language Models (LLMs) still struggle with the ``lost-in-the-middle'' problem, where critical information located in the middle of long-context inputs is often underrepresented or lost. While existing methods attempt to address this by combining multi-scale rotary position embeddings (RoPE), they typically suffer from high latency or rely on suboptimal hand-crafted scaling strategies. To overcome these limitations, we introduce a layer-specific positional embedding scaling~(LPES) method that assigns distinct scaling factors to each layer. LPES achieves a more balanced attention distribut"},"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":"2606.27705","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-26T04:07:41Z","cross_cats_sorted":[],"title_canon_sha256":"9b016afecb2d64c3d9b60bf98c31a8129854826da07bb8a15df76a7c7e5a9bc6","abstract_canon_sha256":"3791b2c3430835e86ace5a9f05fc116484975429d7a5de264418607bde94e9dc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-29T01:14:45.998776Z","signature_b64":"QwmVu+AR42SeruNVqkWlExEw6MFDaPdsJSyxe19pbRxFhgLaweqeLoow+99whdU58K0GNAWZsRyiowhdHPzaBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d12d381d38c1322b3742730f7ed8e3173d6588eefe31a6ae4f61e28e706add79","last_reissued_at":"2026-06-29T01:14:45.998406Z","signature_status":"signed_v1","first_computed_at":"2026-06-29T01:14:45.998406Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Mitigating Position Bias in Transformers via Layer-Specific Positional Embedding Scaling","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Changze Lv, Muling Wu, Qi Qian, Shizheng Li, Tianlong Li, Tianyuan Shi, Xiaoqing Zheng, Xuanjing Huang, Yiran Ding, Yixin Wu, Zhenghua Wang, Zhibo Xu","submitted_at":"2026-06-26T04:07:41Z","abstract_excerpt":"Large Language Models (LLMs) still struggle with the ``lost-in-the-middle'' problem, where critical information located in the middle of long-context inputs is often underrepresented or lost. While existing methods attempt to address this by combining multi-scale rotary position embeddings (RoPE), they typically suffer from high latency or rely on suboptimal hand-crafted scaling strategies. To overcome these limitations, we introduce a layer-specific positional embedding scaling~(LPES) method that assigns distinct scaling factors to each layer. LPES achieves a more balanced attention distribut"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.27705","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/2606.27705/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":"2606.27705","created_at":"2026-06-29T01:14:45.998465+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.27705v1","created_at":"2026-06-29T01:14:45.998465+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.27705","created_at":"2026-06-29T01:14:45.998465+00:00"},{"alias_kind":"pith_short_12","alias_value":"2EWTQHJYYEZC","created_at":"2026-06-29T01:14:45.998465+00:00"},{"alias_kind":"pith_short_16","alias_value":"2EWTQHJYYEZCWN2C","created_at":"2026-06-29T01:14:45.998465+00:00"},{"alias_kind":"pith_short_8","alias_value":"2EWTQHJY","created_at":"2026-06-29T01:14:45.998465+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/2EWTQHJYYEZCWN2COMHX5WHDC4","json":"https://pith.science/pith/2EWTQHJYYEZCWN2COMHX5WHDC4.json","graph_json":"https://pith.science/api/pith-number/2EWTQHJYYEZCWN2COMHX5WHDC4/graph.json","events_json":"https://pith.science/api/pith-number/2EWTQHJYYEZCWN2COMHX5WHDC4/events.json","paper":"https://pith.science/paper/2EWTQHJY"},"agent_actions":{"view_html":"https://pith.science/pith/2EWTQHJYYEZCWN2COMHX5WHDC4","download_json":"https://pith.science/pith/2EWTQHJYYEZCWN2COMHX5WHDC4.json","view_paper":"https://pith.science/paper/2EWTQHJY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.27705&json=true","fetch_graph":"https://pith.science/api/pith-number/2EWTQHJYYEZCWN2COMHX5WHDC4/graph.json","fetch_events":"https://pith.science/api/pith-number/2EWTQHJYYEZCWN2COMHX5WHDC4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2EWTQHJYYEZCWN2COMHX5WHDC4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2EWTQHJYYEZCWN2COMHX5WHDC4/action/storage_attestation","attest_author":"https://pith.science/pith/2EWTQHJYYEZCWN2COMHX5WHDC4/action/author_attestation","sign_citation":"https://pith.science/pith/2EWTQHJYYEZCWN2COMHX5WHDC4/action/citation_signature","submit_replication":"https://pith.science/pith/2EWTQHJYYEZCWN2COMHX5WHDC4/action/replication_record"}},"created_at":"2026-06-29T01:14:45.998465+00:00","updated_at":"2026-06-29T01:14:45.998465+00:00"}