{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:73JJAEJU76TILPW4665O4KPY5H","short_pith_number":"pith:73JJAEJU","schema_version":"1.0","canonical_sha256":"fed2901134ffa685bedcf7baee29f8e9fd735b2af1b6af67d0b10b045e4a47b3","source":{"kind":"arxiv","id":"2305.14628","version":2},"attestation_state":"computed","paper":{"title":"Getting MoRE out of Mixture of Language Model Reasoning Experts","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chenglei Si, Chen Zhao, Jordan Boyd-Graber, Luke Zettlemoyer, Weijia Shi","submitted_at":"2023-05-24T02:00:51Z","abstract_excerpt":"While recent large language models (LLMs) improve on various question answering (QA) datasets, it remains difficult for a single model to generalize across question types that require distinct reasoning abilities. We provide empirical evidence that state-of-the-art LLMs suffer from poor generalizability on reasoning types beyond those seen in the prompt. To remedy this, we propose a Mixture-of-Reasoning-Experts (MoRE) framework that ensembles diverse specialized language models. We specialize the backbone language model with prompts optimized for different reasoning categories, including factu"},"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":"2305.14628","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-05-24T02:00:51Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ad24e0a63df029aa8aea1ea8fce23c0728c137785ed8da1c6cc59d09afcf5f98","abstract_canon_sha256":"9c7c7aa1a15ded5fc59851a5506b4e70030d132a4f0c3a164d49d3482998c369"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:02:50.687608Z","signature_b64":"dZvdB33Iuv+R4n7kYbLsbFdPtqVqFaqmNW2gsaTgRJY6E/YsefAvsVDlFF6jZ87cX7wNxErM50OEZksZ5RT+CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fed2901134ffa685bedcf7baee29f8e9fd735b2af1b6af67d0b10b045e4a47b3","last_reissued_at":"2026-07-05T07:02:50.687019Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:02:50.687019Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Getting MoRE out of Mixture of Language Model Reasoning Experts","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chenglei Si, Chen Zhao, Jordan Boyd-Graber, Luke Zettlemoyer, Weijia Shi","submitted_at":"2023-05-24T02:00:51Z","abstract_excerpt":"While recent large language models (LLMs) improve on various question answering (QA) datasets, it remains difficult for a single model to generalize across question types that require distinct reasoning abilities. We provide empirical evidence that state-of-the-art LLMs suffer from poor generalizability on reasoning types beyond those seen in the prompt. To remedy this, we propose a Mixture-of-Reasoning-Experts (MoRE) framework that ensembles diverse specialized language models. We specialize the backbone language model with prompts optimized for different reasoning categories, including factu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.14628","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/2305.14628/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":"2305.14628","created_at":"2026-07-05T07:02:50.687089+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.14628v2","created_at":"2026-07-05T07:02:50.687089+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.14628","created_at":"2026-07-05T07:02:50.687089+00:00"},{"alias_kind":"pith_short_12","alias_value":"73JJAEJU76TI","created_at":"2026-07-05T07:02:50.687089+00:00"},{"alias_kind":"pith_short_16","alias_value":"73JJAEJU76TILPW4","created_at":"2026-07-05T07:02:50.687089+00:00"},{"alias_kind":"pith_short_8","alias_value":"73JJAEJU","created_at":"2026-07-05T07:02:50.687089+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.00519","citing_title":"CausalAbstain: Enhancing Multilingual LLMs with Causal Reasoning for Trustworthy Abstention","ref_index":31,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/73JJAEJU76TILPW4665O4KPY5H","json":"https://pith.science/pith/73JJAEJU76TILPW4665O4KPY5H.json","graph_json":"https://pith.science/api/pith-number/73JJAEJU76TILPW4665O4KPY5H/graph.json","events_json":"https://pith.science/api/pith-number/73JJAEJU76TILPW4665O4KPY5H/events.json","paper":"https://pith.science/paper/73JJAEJU"},"agent_actions":{"view_html":"https://pith.science/pith/73JJAEJU76TILPW4665O4KPY5H","download_json":"https://pith.science/pith/73JJAEJU76TILPW4665O4KPY5H.json","view_paper":"https://pith.science/paper/73JJAEJU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.14628&json=true","fetch_graph":"https://pith.science/api/pith-number/73JJAEJU76TILPW4665O4KPY5H/graph.json","fetch_events":"https://pith.science/api/pith-number/73JJAEJU76TILPW4665O4KPY5H/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/73JJAEJU76TILPW4665O4KPY5H/action/timestamp_anchor","attest_storage":"https://pith.science/pith/73JJAEJU76TILPW4665O4KPY5H/action/storage_attestation","attest_author":"https://pith.science/pith/73JJAEJU76TILPW4665O4KPY5H/action/author_attestation","sign_citation":"https://pith.science/pith/73JJAEJU76TILPW4665O4KPY5H/action/citation_signature","submit_replication":"https://pith.science/pith/73JJAEJU76TILPW4665O4KPY5H/action/replication_record"}},"created_at":"2026-07-05T07:02:50.687089+00:00","updated_at":"2026-07-05T07:02:50.687089+00:00"}