{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:6P2E2DW3LAMIBVGHRQZQS5QI7J","short_pith_number":"pith:6P2E2DW3","schema_version":"1.0","canonical_sha256":"f3f44d0edb581880d4c78c33097608fa690b7bbaa68e52ee26a79bc31b3acd2e","source":{"kind":"arxiv","id":"2502.05167","version":3},"attestation_state":"computed","paper":{"title":"NoLiMa: Long-Context Evaluation Beyond Literal Matching","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Ali Modarressi, Franck Dernoncourt, Hanieh Deilamsalehy, Hinrich Sch\\\"utze, Ryan A. Rossi, Seunghyun Yoon, Trung Bui","submitted_at":"2025-02-07T18:49:46Z","abstract_excerpt":"Recent large language models (LLMs) support long contexts ranging from 128K to 1M tokens. A popular method for evaluating these capabilities is the needle-in-a-haystack (NIAH) test, which involves retrieving a \"needle\" (relevant information) from a \"haystack\" (long irrelevant context). Extensions of this approach include increasing distractors, fact chaining, and in-context reasoning. However, in these benchmarks, models can exploit existing literal matches between the needle and haystack to simplify the task. To address this, we introduce NoLiMa, a benchmark extending NIAH with a carefully de"},"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":"2502.05167","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-07T18:49:46Z","cross_cats_sorted":[],"title_canon_sha256":"986acf6d4fbfc2fdfde2c45a35e4b9efaf99ed304e73678c5c725eb589e4bab8","abstract_canon_sha256":"3c50f4295312b4bb3f5d5e99c5c9bc0eb410e105d6f08718b23526dc80a60d2c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:34:04.298877Z","signature_b64":"VQiC1KinjQW1s7g4vuDdAG86J8sXF9v3UN/UkIhsNXdZ7j70siKyiPmyEhAHrVvTpA2M6uuKmzh4U+MB3SEnDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f3f44d0edb581880d4c78c33097608fa690b7bbaa68e52ee26a79bc31b3acd2e","last_reissued_at":"2026-07-05T11:34:04.298389Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:34:04.298389Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"NoLiMa: Long-Context Evaluation Beyond Literal Matching","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Ali Modarressi, Franck Dernoncourt, Hanieh Deilamsalehy, Hinrich Sch\\\"utze, Ryan A. Rossi, Seunghyun Yoon, Trung Bui","submitted_at":"2025-02-07T18:49:46Z","abstract_excerpt":"Recent large language models (LLMs) support long contexts ranging from 128K to 1M tokens. A popular method for evaluating these capabilities is the needle-in-a-haystack (NIAH) test, which involves retrieving a \"needle\" (relevant information) from a \"haystack\" (long irrelevant context). Extensions of this approach include increasing distractors, fact chaining, and in-context reasoning. However, in these benchmarks, models can exploit existing literal matches between the needle and haystack to simplify the task. To address this, we introduce NoLiMa, a benchmark extending NIAH with a carefully de"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.05167","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/2502.05167/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":"2502.05167","created_at":"2026-07-05T11:34:04.298454+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.05167v3","created_at":"2026-07-05T11:34:04.298454+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.05167","created_at":"2026-07-05T11:34:04.298454+00:00"},{"alias_kind":"pith_short_12","alias_value":"6P2E2DW3LAMI","created_at":"2026-07-05T11:34:04.298454+00:00"},{"alias_kind":"pith_short_16","alias_value":"6P2E2DW3LAMIBVGH","created_at":"2026-07-05T11:34:04.298454+00:00"},{"alias_kind":"pith_short_8","alias_value":"6P2E2DW3","created_at":"2026-07-05T11:34:04.298454+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":19,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.08032","citing_title":"What to Keep, What to Forget: A Rate--Distortion View of Memory Compaction in LLMs and Agents","ref_index":82,"is_internal_anchor":true},{"citing_arxiv_id":"2606.06906","citing_title":"EASE-TTT: Evidence-Aligned Selective Test-Time Training for Long-Context Question Answering","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2606.04660","citing_title":"LifeSide: Benchmarking Agents as Lifelong Digital Companions","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02963","citing_title":"KForge: LLM-Driven Cross-Platform Kernel Generation for AI Accelerators","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2605.24279","citing_title":"ContextEcho: A Benchmark for Persona Drift in Long Agentic-Coding Sessions","ref_index":56,"is_internal_anchor":false},{"citing_arxiv_id":"2605.27294","citing_title":"Separating Semantic Competition from Context Length in RAG Reading","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2606.06840","citing_title":"Characterize Then Distill: Mechanistic Reasoning in Large Output Spaces","ref_index":154,"is_internal_anchor":false},{"citing_arxiv_id":"2605.23170","citing_title":"Positional Failures in Long-Context LLMs: A Blind Spot in Reasoning Benchmarks","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2605.23296","citing_title":"Parallel Context Compaction for Long-Horizon LLM Agent Serving","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2601.21468","citing_title":"MemOCR: Layout-Aware Visual Memory for Efficient Long-Horizon Reasoning","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2507.05257","citing_title":"Evaluating Memory in LLM Agents via Incremental Multi-Turn Interactions","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2603.13091","citing_title":"Reasoning over Video: Evaluating How MLLMs Extract, Integrate, and Reconstruct Spatiotemporal Evidence","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2601.02780","citing_title":"MiMo-V2-Flash Technical Report","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10544","citing_title":"Where Does Long-Context Supervision Actually Go? Effective-Context Exposure Balancing","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08580","citing_title":"Slipstream: Trajectory-Grounded Compaction Validation for Long-Horizon Agents","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05806","citing_title":"Retrieval from Within: An Intrinsic Capability of Attention-Based Models","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2604.21816","citing_title":"Tool Attention Is All You Need: Dynamic Tool Gating and Lazy Schema Loading for Eliminating the MCP/Tools Tax in Scalable Agentic Workflows","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05806","citing_title":"Retrieval from Within: An Intrinsic Capability of Attention-Based Models","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2604.05151","citing_title":"Context Collapse: Barriers to Adoption for Generative AI in Workplace Settings","ref_index":89,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6P2E2DW3LAMIBVGHRQZQS5QI7J","json":"https://pith.science/pith/6P2E2DW3LAMIBVGHRQZQS5QI7J.json","graph_json":"https://pith.science/api/pith-number/6P2E2DW3LAMIBVGHRQZQS5QI7J/graph.json","events_json":"https://pith.science/api/pith-number/6P2E2DW3LAMIBVGHRQZQS5QI7J/events.json","paper":"https://pith.science/paper/6P2E2DW3"},"agent_actions":{"view_html":"https://pith.science/pith/6P2E2DW3LAMIBVGHRQZQS5QI7J","download_json":"https://pith.science/pith/6P2E2DW3LAMIBVGHRQZQS5QI7J.json","view_paper":"https://pith.science/paper/6P2E2DW3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.05167&json=true","fetch_graph":"https://pith.science/api/pith-number/6P2E2DW3LAMIBVGHRQZQS5QI7J/graph.json","fetch_events":"https://pith.science/api/pith-number/6P2E2DW3LAMIBVGHRQZQS5QI7J/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6P2E2DW3LAMIBVGHRQZQS5QI7J/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6P2E2DW3LAMIBVGHRQZQS5QI7J/action/storage_attestation","attest_author":"https://pith.science/pith/6P2E2DW3LAMIBVGHRQZQS5QI7J/action/author_attestation","sign_citation":"https://pith.science/pith/6P2E2DW3LAMIBVGHRQZQS5QI7J/action/citation_signature","submit_replication":"https://pith.science/pith/6P2E2DW3LAMIBVGHRQZQS5QI7J/action/replication_record"}},"created_at":"2026-07-05T11:34:04.298454+00:00","updated_at":"2026-07-05T11:34:04.298454+00:00"}