{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:CXCQB6TEKEYYBCT2LYD2YZDR5D","short_pith_number":"pith:CXCQB6TE","schema_version":"1.0","canonical_sha256":"15c500fa645131808a7a5e07ac6471e8d71f83f19ea92be8ad8f3cda1d431e5f","source":{"kind":"arxiv","id":"2408.01866","version":3},"attestation_state":"computed","paper":{"title":"Efficient Solutions For An Intriguing Failure of LLMs: Long Context Window Does Not Mean LLMs Can Analyze Long Sequences Flawlessly","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Iacopo Ghinassi, Ignacio Castro, Matthew Purver, Peyman Hosseini","submitted_at":"2024-08-03T21:31:34Z","abstract_excerpt":"Large Language Models (LLMs) have demonstrated remarkable capabilities in comprehending and analyzing lengthy sequential inputs, owing to their extensive context windows that allow processing millions of tokens in a single forward pass. However, this paper uncovers a surprising limitation: LLMs fall short when handling long input sequences. We investigate this issue using three datasets and two tasks (sentiment analysis and news categorization) across various LLMs, including Claude 3, Gemini Pro, GPT 3.5 Turbo, Llama 3 Instruct, and Mistral Instruct models. To address this limitation, we propo"},"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":"2408.01866","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-08-03T21:31:34Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"93fe4637fbc3490b3d7198fda45484d6a9967455a58e46201f611e849a1a1883","abstract_canon_sha256":"b387b3052e4071565e141d473f7bbd2eb0a72ab4176c399289e3b29f6c317362"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:52:14.358244Z","signature_b64":"1aqxHpoROo4R/12CnC2iZFNs4R1v5gFoANehfTHrGbC8PYxINT182vl9JvLJUa2y5cDSkrOR186KYZm2wNS7BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"15c500fa645131808a7a5e07ac6471e8d71f83f19ea92be8ad8f3cda1d431e5f","last_reissued_at":"2026-07-05T09:52:14.357797Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:52:14.357797Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Efficient Solutions For An Intriguing Failure of LLMs: Long Context Window Does Not Mean LLMs Can Analyze Long Sequences Flawlessly","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Iacopo Ghinassi, Ignacio Castro, Matthew Purver, Peyman Hosseini","submitted_at":"2024-08-03T21:31:34Z","abstract_excerpt":"Large Language Models (LLMs) have demonstrated remarkable capabilities in comprehending and analyzing lengthy sequential inputs, owing to their extensive context windows that allow processing millions of tokens in a single forward pass. However, this paper uncovers a surprising limitation: LLMs fall short when handling long input sequences. We investigate this issue using three datasets and two tasks (sentiment analysis and news categorization) across various LLMs, including Claude 3, Gemini Pro, GPT 3.5 Turbo, Llama 3 Instruct, and Mistral Instruct models. To address this limitation, we propo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.01866","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/2408.01866/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":"2408.01866","created_at":"2026-07-05T09:52:14.357859+00:00"},{"alias_kind":"arxiv_version","alias_value":"2408.01866v3","created_at":"2026-07-05T09:52:14.357859+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.01866","created_at":"2026-07-05T09:52:14.357859+00:00"},{"alias_kind":"pith_short_12","alias_value":"CXCQB6TEKEYY","created_at":"2026-07-05T09:52:14.357859+00:00"},{"alias_kind":"pith_short_16","alias_value":"CXCQB6TEKEYYBCT2","created_at":"2026-07-05T09:52:14.357859+00:00"},{"alias_kind":"pith_short_8","alias_value":"CXCQB6TE","created_at":"2026-07-05T09:52:14.357859+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.06566","citing_title":"NTILC: Neural Tool Invocation via Learned Compression","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2605.29368","citing_title":"SURGENT: A Surgical Multi-Agent Assistance System Across the Perioperative Workflow","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2505.19353","citing_title":"Architectures of Error: A Philosophical Inquiry into AI and Human Code Generation","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06330","citing_title":"Fine-Tuning Small Language Models for Solution-Oriented Windows Event Log Analysis","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2604.08851","citing_title":"Cross-Lingual Attention Distillation with Personality-Informed Generative Augmentation for Multilingual Personality Recognition","ref_index":8,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CXCQB6TEKEYYBCT2LYD2YZDR5D","json":"https://pith.science/pith/CXCQB6TEKEYYBCT2LYD2YZDR5D.json","graph_json":"https://pith.science/api/pith-number/CXCQB6TEKEYYBCT2LYD2YZDR5D/graph.json","events_json":"https://pith.science/api/pith-number/CXCQB6TEKEYYBCT2LYD2YZDR5D/events.json","paper":"https://pith.science/paper/CXCQB6TE"},"agent_actions":{"view_html":"https://pith.science/pith/CXCQB6TEKEYYBCT2LYD2YZDR5D","download_json":"https://pith.science/pith/CXCQB6TEKEYYBCT2LYD2YZDR5D.json","view_paper":"https://pith.science/paper/CXCQB6TE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2408.01866&json=true","fetch_graph":"https://pith.science/api/pith-number/CXCQB6TEKEYYBCT2LYD2YZDR5D/graph.json","fetch_events":"https://pith.science/api/pith-number/CXCQB6TEKEYYBCT2LYD2YZDR5D/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CXCQB6TEKEYYBCT2LYD2YZDR5D/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CXCQB6TEKEYYBCT2LYD2YZDR5D/action/storage_attestation","attest_author":"https://pith.science/pith/CXCQB6TEKEYYBCT2LYD2YZDR5D/action/author_attestation","sign_citation":"https://pith.science/pith/CXCQB6TEKEYYBCT2LYD2YZDR5D/action/citation_signature","submit_replication":"https://pith.science/pith/CXCQB6TEKEYYBCT2LYD2YZDR5D/action/replication_record"}},"created_at":"2026-07-05T09:52:14.357859+00:00","updated_at":"2026-07-05T09:52:14.357859+00:00"}