{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:MRZWJWMM6VCWJUGH5ARVCOKJ6I","short_pith_number":"pith:MRZWJWMM","schema_version":"1.0","canonical_sha256":"647364d98cf54564d0c7e823513949f21e99ab6110d159e6ea1588775baef6bb","source":{"kind":"arxiv","id":"2502.17125","version":1},"attestation_state":"computed","paper":{"title":"LettuceDetect: A Hallucination Detection Framework for RAG Applications","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"\\'Ad\\'am Kov\\'acs, G\\'abor Recski","submitted_at":"2025-02-24T13:11:47Z","abstract_excerpt":"Retrieval Augmented Generation (RAG) systems remain vulnerable to hallucinated answers despite incorporating external knowledge sources. We present LettuceDetect a framework that addresses two critical limitations in existing hallucination detection methods: (1) the context window constraints of traditional encoder-based methods, and (2) the computational inefficiency of LLM based approaches. Building on ModernBERT's extended context capabilities (up to 8k tokens) and trained on the RAGTruth benchmark dataset, our approach outperforms all previous encoder-based models and most prompt-based mod"},"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":"2502.17125","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-24T13:11:47Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"9658a3356cb372bf83e74e848596ccded260c968cb2256b12e35a96007e35c87","abstract_canon_sha256":"4efc75b6dd9dfc20eb796a2a78179033d45e62ba05d3707a2b977e59f89814d8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:19:06.329345Z","signature_b64":"OWtCFpXl4YyFvtX0owUPcQxR8lmGJXg2E8GlpmwhyIFcnTmr4YRryDOv3vDblgeB5M7NrgFzE5pvcqKOvPzFCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"647364d98cf54564d0c7e823513949f21e99ab6110d159e6ea1588775baef6bb","last_reissued_at":"2026-07-05T10:19:06.328861Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:19:06.328861Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LettuceDetect: A Hallucination Detection Framework for RAG Applications","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"\\'Ad\\'am Kov\\'acs, G\\'abor Recski","submitted_at":"2025-02-24T13:11:47Z","abstract_excerpt":"Retrieval Augmented Generation (RAG) systems remain vulnerable to hallucinated answers despite incorporating external knowledge sources. We present LettuceDetect a framework that addresses two critical limitations in existing hallucination detection methods: (1) the context window constraints of traditional encoder-based methods, and (2) the computational inefficiency of LLM based approaches. Building on ModernBERT's extended context capabilities (up to 8k tokens) and trained on the RAGTruth benchmark dataset, our approach outperforms all previous encoder-based models and most prompt-based mod"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.17125","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/2502.17125/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":"2502.17125","created_at":"2026-07-05T10:19:06.328916+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.17125v1","created_at":"2026-07-05T10:19:06.328916+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.17125","created_at":"2026-07-05T10:19:06.328916+00:00"},{"alias_kind":"pith_short_12","alias_value":"MRZWJWMM6VCW","created_at":"2026-07-05T10:19:06.328916+00:00"},{"alias_kind":"pith_short_16","alias_value":"MRZWJWMM6VCWJUGH","created_at":"2026-07-05T10:19:06.328916+00:00"},{"alias_kind":"pith_short_8","alias_value":"MRZWJWMM","created_at":"2026-07-05T10:19:06.328916+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.19351","citing_title":"Detecting Hallucinations for Large Language Model-based Knowledge Graph Reasoning","ref_index":108,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08462","citing_title":"Do Benchmarks Underestimate LLM Performance? Evaluating Hallucination Detection With LLM-First Human-Adjudicated Assessment","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2604.15945","citing_title":"RAGognizer: Hallucination-Aware Fine-Tuning via Detection Head Integration","ref_index":39,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MRZWJWMM6VCWJUGH5ARVCOKJ6I","json":"https://pith.science/pith/MRZWJWMM6VCWJUGH5ARVCOKJ6I.json","graph_json":"https://pith.science/api/pith-number/MRZWJWMM6VCWJUGH5ARVCOKJ6I/graph.json","events_json":"https://pith.science/api/pith-number/MRZWJWMM6VCWJUGH5ARVCOKJ6I/events.json","paper":"https://pith.science/paper/MRZWJWMM"},"agent_actions":{"view_html":"https://pith.science/pith/MRZWJWMM6VCWJUGH5ARVCOKJ6I","download_json":"https://pith.science/pith/MRZWJWMM6VCWJUGH5ARVCOKJ6I.json","view_paper":"https://pith.science/paper/MRZWJWMM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.17125&json=true","fetch_graph":"https://pith.science/api/pith-number/MRZWJWMM6VCWJUGH5ARVCOKJ6I/graph.json","fetch_events":"https://pith.science/api/pith-number/MRZWJWMM6VCWJUGH5ARVCOKJ6I/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MRZWJWMM6VCWJUGH5ARVCOKJ6I/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MRZWJWMM6VCWJUGH5ARVCOKJ6I/action/storage_attestation","attest_author":"https://pith.science/pith/MRZWJWMM6VCWJUGH5ARVCOKJ6I/action/author_attestation","sign_citation":"https://pith.science/pith/MRZWJWMM6VCWJUGH5ARVCOKJ6I/action/citation_signature","submit_replication":"https://pith.science/pith/MRZWJWMM6VCWJUGH5ARVCOKJ6I/action/replication_record"}},"created_at":"2026-07-05T10:19:06.328916+00:00","updated_at":"2026-07-05T10:19:06.328916+00:00"}