{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:QXWF3YEINF7BBAGIVKOOK64THL","short_pith_number":"pith:QXWF3YEI","schema_version":"1.0","canonical_sha256":"85ec5de088697e1080c8aa9ce57b933aed9e70199b2aaab07286cdd88b163c19","source":{"kind":"arxiv","id":"2502.18915","version":3},"attestation_state":"computed","paper":{"title":"END: Early Noise Dropping for Efficient and Effective Context Denoising","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Bing Yin, Binxuan Huang, Fangran Mo, Hongye Jin, Huasheng Li, Jingfeng Yang, Jinghan Zhang, Meng Jiang, Pei Chen, Tianyi Liu, Xinyang Zhang, Yifan Gao, Zheng Li, Zhengyang Wang","submitted_at":"2025-02-26T08:07:17Z","abstract_excerpt":"Large Language Models (LLMs) have demonstrated remarkable performance across a wide range of natural language processing tasks. However, they are often distracted by irrelevant or noisy context in input sequences that degrades output quality. This problem affects both long- and short-context scenarios, such as retrieval-augmented generation, table question-answering, and in-context learning. We reveal that LLMs can implicitly identify whether input sequences contain useful information at early layers, prior to token generation. Leveraging this insight, we introduce Early Noise Dropping (\\texts"},"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":"2502.18915","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-26T08:07:17Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"6568a6a187ba14cedcbbba72686a8152db4b994dc0ca572df2eec0c8a257a48d","abstract_canon_sha256":"6c7d6c43d0a4320ce445489181cb3e97a39a90ae4d2a5346310388c8e1f7ff15"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T01:04:13.141471Z","signature_b64":"QxKQw6zQsvWGxf2JSpqcNN4vJrSKk+/Wg+pPLOvXEUoL8HTY9IboQJjFtXkMV1LHJiG0/c8O3UfYh5WEUoW5BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"85ec5de088697e1080c8aa9ce57b933aed9e70199b2aaab07286cdd88b163c19","last_reissued_at":"2026-05-21T01:04:13.140875Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T01:04:13.140875Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"END: Early Noise Dropping for Efficient and Effective Context Denoising","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Bing Yin, Binxuan Huang, Fangran Mo, Hongye Jin, Huasheng Li, Jingfeng Yang, Jinghan Zhang, Meng Jiang, Pei Chen, Tianyi Liu, Xinyang Zhang, Yifan Gao, Zheng Li, Zhengyang Wang","submitted_at":"2025-02-26T08:07:17Z","abstract_excerpt":"Large Language Models (LLMs) have demonstrated remarkable performance across a wide range of natural language processing tasks. However, they are often distracted by irrelevant or noisy context in input sequences that degrades output quality. This problem affects both long- and short-context scenarios, such as retrieval-augmented generation, table question-answering, and in-context learning. We reveal that LLMs can implicitly identify whether input sequences contain useful information at early layers, prior to token generation. Leveraging this insight, we introduce Early Noise Dropping (\\texts"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.18915","kind":"arxiv","version":3},"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/2502.18915/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":"2502.18915","created_at":"2026-05-21T01:04:13.140983+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.18915v3","created_at":"2026-05-21T01:04:13.140983+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.18915","created_at":"2026-05-21T01:04:13.140983+00:00"},{"alias_kind":"pith_short_12","alias_value":"QXWF3YEINF7B","created_at":"2026-05-21T01:04:13.140983+00:00"},{"alias_kind":"pith_short_16","alias_value":"QXWF3YEINF7BBAGI","created_at":"2026-05-21T01:04:13.140983+00:00"},{"alias_kind":"pith_short_8","alias_value":"QXWF3YEI","created_at":"2026-05-21T01:04:13.140983+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/QXWF3YEINF7BBAGIVKOOK64THL","json":"https://pith.science/pith/QXWF3YEINF7BBAGIVKOOK64THL.json","graph_json":"https://pith.science/api/pith-number/QXWF3YEINF7BBAGIVKOOK64THL/graph.json","events_json":"https://pith.science/api/pith-number/QXWF3YEINF7BBAGIVKOOK64THL/events.json","paper":"https://pith.science/paper/QXWF3YEI"},"agent_actions":{"view_html":"https://pith.science/pith/QXWF3YEINF7BBAGIVKOOK64THL","download_json":"https://pith.science/pith/QXWF3YEINF7BBAGIVKOOK64THL.json","view_paper":"https://pith.science/paper/QXWF3YEI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.18915&json=true","fetch_graph":"https://pith.science/api/pith-number/QXWF3YEINF7BBAGIVKOOK64THL/graph.json","fetch_events":"https://pith.science/api/pith-number/QXWF3YEINF7BBAGIVKOOK64THL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QXWF3YEINF7BBAGIVKOOK64THL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QXWF3YEINF7BBAGIVKOOK64THL/action/storage_attestation","attest_author":"https://pith.science/pith/QXWF3YEINF7BBAGIVKOOK64THL/action/author_attestation","sign_citation":"https://pith.science/pith/QXWF3YEINF7BBAGIVKOOK64THL/action/citation_signature","submit_replication":"https://pith.science/pith/QXWF3YEINF7BBAGIVKOOK64THL/action/replication_record"}},"created_at":"2026-05-21T01:04:13.140983+00:00","updated_at":"2026-05-21T01:04:13.140983+00:00"}