{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:BJ36JCEJGPFFSN3YQTJR6JSEUQ","short_pith_number":"pith:BJ36JCEJ","schema_version":"1.0","canonical_sha256":"0a77e4888933ca59377884d31f2644a42b92535b80c602d4819b9826a07fb418","source":{"kind":"arxiv","id":"2402.04617","version":2},"attestation_state":"computed","paper":{"title":"InfLLM: Training-Free Long-Context Extrapolation for LLMs with an Efficient Context Memory","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Chaojun Xiao, Guangxuan Xiao, Maosong Sun, Pengle Zhang, Xu Han, Yankai Lin, Zhengyan Zhang, Zhiyuan Liu","submitted_at":"2024-02-07T06:50:42Z","abstract_excerpt":"Large language models (LLMs) have emerged as a cornerstone in real-world applications with lengthy streaming inputs (e.g., LLM-driven agents). However, existing LLMs, pre-trained on sequences with a restricted maximum length, cannot process longer sequences due to the out-of-domain and distraction issues. Common solutions often involve continual pre-training on longer sequences, which will introduce expensive computational overhead and uncontrollable change in model capabilities. In this paper, we unveil the intrinsic capacity of LLMs for understanding extremely long sequences without any fine"},"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":"2402.04617","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-07T06:50:42Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"be61b630fc30b72d6e5ecf9943125be4a1485127958563b123f0a48efe3f90aa","abstract_canon_sha256":"e6f39c18d89f4575fe0fa59a8447505255551caf12d1b99dbd9ca84e015d7d6a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:24:01.177445Z","signature_b64":"10vppD11G9MhEhhxHLwNSSwcWSbOK3rmRV1SDebnBTLJZyXnnFFQ1i2MtPXgXob1gvSHVwxb7//ymGKkH3eWAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0a77e4888933ca59377884d31f2644a42b92535b80c602d4819b9826a07fb418","last_reissued_at":"2026-07-05T08:24:01.176992Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:24:01.176992Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"InfLLM: Training-Free Long-Context Extrapolation for LLMs with an Efficient Context Memory","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Chaojun Xiao, Guangxuan Xiao, Maosong Sun, Pengle Zhang, Xu Han, Yankai Lin, Zhengyan Zhang, Zhiyuan Liu","submitted_at":"2024-02-07T06:50:42Z","abstract_excerpt":"Large language models (LLMs) have emerged as a cornerstone in real-world applications with lengthy streaming inputs (e.g., LLM-driven agents). However, existing LLMs, pre-trained on sequences with a restricted maximum length, cannot process longer sequences due to the out-of-domain and distraction issues. Common solutions often involve continual pre-training on longer sequences, which will introduce expensive computational overhead and uncontrollable change in model capabilities. In this paper, we unveil the intrinsic capacity of LLMs for understanding extremely long sequences without any fine"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.04617","kind":"arxiv","version":2},"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/2402.04617/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":"2402.04617","created_at":"2026-07-05T08:24:01.177049+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.04617v2","created_at":"2026-07-05T08:24:01.177049+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.04617","created_at":"2026-07-05T08:24:01.177049+00:00"},{"alias_kind":"pith_short_12","alias_value":"BJ36JCEJGPFF","created_at":"2026-07-05T08:24:01.177049+00:00"},{"alias_kind":"pith_short_16","alias_value":"BJ36JCEJGPFFSN3Y","created_at":"2026-07-05T08:24:01.177049+00:00"},{"alias_kind":"pith_short_8","alias_value":"BJ36JCEJ","created_at":"2026-07-05T08:24:01.177049+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":10,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2503.19950","citing_title":"LogQuant: Log-Distributed 2-Bit Quantization of KV Cache with Superior Accuracy Preservation","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2404.07143","citing_title":"Leave No Context Behind: Efficient Infinite Context Transformers with Infini-attention","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18053","citing_title":"Protection Is (Nearly) All You Need: Structural Protection Dominates Scoring in Globally Capped KV Eviction","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2409.10516","citing_title":"RetrievalAttention: Accelerating Long-Context LLM Inference via Vector Retrieval","ref_index":114,"is_internal_anchor":false},{"citing_arxiv_id":"2601.11913","citing_title":"LSTM-MAS: A Long Short-Term Memory Inspired Multi-Agent System for Long-Context Understanding","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2603.04759","citing_title":"Stacked from One: Multi-Scale Self-Injection for Context Window Extension","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2604.03263","citing_title":"LPC-SM: Local Predictive Coding and Sparse Memory for Long-Context Language Modeling","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13831","citing_title":"Training Long-Context Vision-Language Models Effectively with Generalization Beyond 128K Context","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2404.06654","citing_title":"RULER: What's the Real Context Size of Your Long-Context Language Models?","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07363","citing_title":"MISA: Mixture of Indexer Sparse Attention for Long-Context LLM Inference","ref_index":22,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BJ36JCEJGPFFSN3YQTJR6JSEUQ","json":"https://pith.science/pith/BJ36JCEJGPFFSN3YQTJR6JSEUQ.json","graph_json":"https://pith.science/api/pith-number/BJ36JCEJGPFFSN3YQTJR6JSEUQ/graph.json","events_json":"https://pith.science/api/pith-number/BJ36JCEJGPFFSN3YQTJR6JSEUQ/events.json","paper":"https://pith.science/paper/BJ36JCEJ"},"agent_actions":{"view_html":"https://pith.science/pith/BJ36JCEJGPFFSN3YQTJR6JSEUQ","download_json":"https://pith.science/pith/BJ36JCEJGPFFSN3YQTJR6JSEUQ.json","view_paper":"https://pith.science/paper/BJ36JCEJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.04617&json=true","fetch_graph":"https://pith.science/api/pith-number/BJ36JCEJGPFFSN3YQTJR6JSEUQ/graph.json","fetch_events":"https://pith.science/api/pith-number/BJ36JCEJGPFFSN3YQTJR6JSEUQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BJ36JCEJGPFFSN3YQTJR6JSEUQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BJ36JCEJGPFFSN3YQTJR6JSEUQ/action/storage_attestation","attest_author":"https://pith.science/pith/BJ36JCEJGPFFSN3YQTJR6JSEUQ/action/author_attestation","sign_citation":"https://pith.science/pith/BJ36JCEJGPFFSN3YQTJR6JSEUQ/action/citation_signature","submit_replication":"https://pith.science/pith/BJ36JCEJGPFFSN3YQTJR6JSEUQ/action/replication_record"}},"created_at":"2026-07-05T08:24:01.177049+00:00","updated_at":"2026-07-05T08:24:01.177049+00:00"}