{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:I23T3BDAZ56GRZAKX3DNIBH5L3","short_pith_number":"pith:I23T3BDA","schema_version":"1.0","canonical_sha256":"46b73d8460cf7c68e40abec6d404fd5ef390050ec201481b82cb4803e46c3d60","source":{"kind":"arxiv","id":"2603.16654","version":2},"attestation_state":"computed","paper":{"title":"Omanic: Towards Step-wise Evaluation of Multi-hop Reasoning in Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Aosong Feng, Chanjun Park, Irene Li, Jinghui Lu, Rex Ying, Sherry T. Tong, Sophia Simeng Han, Xiaojie Gu, Yingjian Chen, Yusuke Iwasawa, Yutaka Matsuo","submitted_at":"2026-03-17T15:23:37Z","abstract_excerpt":"Evaluating the reasoning abilities of large language models (LLMs) solely from final answers can obscure failures in intermediate steps, especially in multi-hop QA benchmarks without step-level annotations. To address this gap, we introduce Omanic, an open-domain 4-hop QA benchmark designed not only to measure final-answer accuracy but also to diagnose where reasoning breaks down. Omanic contains 10,296 machine-generated training examples (OmanicSynth) and 967 expert-reviewed human-annotated evaluation examples (OmanicBench), with each evaluation question decomposed into single-hop sub-questio"},"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":"2603.16654","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-03-17T15:23:37Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"2df7f337d97436fc540018705eaf16da8215be1cecbc05ad2b4282a15bb4d99d","abstract_canon_sha256":"6173ddb2b9f99cf49f42200657462dc15f59331815f4bf7bad4ce08ffb491419"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-27T02:05:13.805305Z","signature_b64":"6rurP/nzmXv3vqOpuHEEGf3TY9jeO6lPcxB2ogvksdd5ITWUHJ/TyI3N5fqNkBOlvcEy6vi7CDKggQDbVKKqCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"46b73d8460cf7c68e40abec6d404fd5ef390050ec201481b82cb4803e46c3d60","last_reissued_at":"2026-05-27T02:05:13.804648Z","signature_status":"signed_v1","first_computed_at":"2026-05-27T02:05:13.804648Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Omanic: Towards Step-wise Evaluation of Multi-hop Reasoning in Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Aosong Feng, Chanjun Park, Irene Li, Jinghui Lu, Rex Ying, Sherry T. Tong, Sophia Simeng Han, Xiaojie Gu, Yingjian Chen, Yusuke Iwasawa, Yutaka Matsuo","submitted_at":"2026-03-17T15:23:37Z","abstract_excerpt":"Evaluating the reasoning abilities of large language models (LLMs) solely from final answers can obscure failures in intermediate steps, especially in multi-hop QA benchmarks without step-level annotations. To address this gap, we introduce Omanic, an open-domain 4-hop QA benchmark designed not only to measure final-answer accuracy but also to diagnose where reasoning breaks down. Omanic contains 10,296 machine-generated training examples (OmanicSynth) and 967 expert-reviewed human-annotated evaluation examples (OmanicBench), with each evaluation question decomposed into single-hop sub-questio"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.16654","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/2603.16654/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":"2603.16654","created_at":"2026-05-27T02:05:13.804746+00:00"},{"alias_kind":"arxiv_version","alias_value":"2603.16654v2","created_at":"2026-05-27T02:05:13.804746+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.16654","created_at":"2026-05-27T02:05:13.804746+00:00"},{"alias_kind":"pith_short_12","alias_value":"I23T3BDAZ56G","created_at":"2026-05-27T02:05:13.804746+00:00"},{"alias_kind":"pith_short_16","alias_value":"I23T3BDAZ56GRZAK","created_at":"2026-05-27T02:05:13.804746+00:00"},{"alias_kind":"pith_short_8","alias_value":"I23T3BDA","created_at":"2026-05-27T02:05:13.804746+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/I23T3BDAZ56GRZAKX3DNIBH5L3","json":"https://pith.science/pith/I23T3BDAZ56GRZAKX3DNIBH5L3.json","graph_json":"https://pith.science/api/pith-number/I23T3BDAZ56GRZAKX3DNIBH5L3/graph.json","events_json":"https://pith.science/api/pith-number/I23T3BDAZ56GRZAKX3DNIBH5L3/events.json","paper":"https://pith.science/paper/I23T3BDA"},"agent_actions":{"view_html":"https://pith.science/pith/I23T3BDAZ56GRZAKX3DNIBH5L3","download_json":"https://pith.science/pith/I23T3BDAZ56GRZAKX3DNIBH5L3.json","view_paper":"https://pith.science/paper/I23T3BDA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2603.16654&json=true","fetch_graph":"https://pith.science/api/pith-number/I23T3BDAZ56GRZAKX3DNIBH5L3/graph.json","fetch_events":"https://pith.science/api/pith-number/I23T3BDAZ56GRZAKX3DNIBH5L3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/I23T3BDAZ56GRZAKX3DNIBH5L3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/I23T3BDAZ56GRZAKX3DNIBH5L3/action/storage_attestation","attest_author":"https://pith.science/pith/I23T3BDAZ56GRZAKX3DNIBH5L3/action/author_attestation","sign_citation":"https://pith.science/pith/I23T3BDAZ56GRZAKX3DNIBH5L3/action/citation_signature","submit_replication":"https://pith.science/pith/I23T3BDAZ56GRZAKX3DNIBH5L3/action/replication_record"}},"created_at":"2026-05-27T02:05:13.804746+00:00","updated_at":"2026-05-27T02:05:13.804746+00:00"}