{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:YA6EVEETUEIHEOEKFUAKNEIKZI","short_pith_number":"pith:YA6EVEET","schema_version":"1.0","canonical_sha256":"c03c4a9093a11072388a2d00a6910aca0c38b9f0002341d0fb0e178c06eb35a3","source":{"kind":"arxiv","id":"2301.12726","version":1},"attestation_state":"computed","paper":{"title":"Specializing Smaller Language Models towards Multi-Step Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Ashish Sabharwal, Hao Peng, Litu Ou, Tushar Khot, Yao Fu","submitted_at":"2023-01-30T08:51:19Z","abstract_excerpt":"The surprising ability of Large Language Models (LLMs) to perform well on complex reasoning with only few-shot chain-of-thought prompts is believed to emerge only in very large-scale models (100+ billion parameters). We show that such abilities can, in fact, be distilled down from GPT-3.5 ($\\ge$ 175B) to T5 variants ($\\le$ 11B). We propose model specialization, to specialize the model's ability towards a target task. The hypothesis is that large models (commonly viewed as larger than 100B) have strong modeling power, but are spread on a large spectrum of tasks. Small models (commonly viewed as"},"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":"2301.12726","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-01-30T08:51:19Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"6ca6e828cde91778fbd6e0968ff811d5846f4e101cbf5ff55485696655787d69","abstract_canon_sha256":"df2307d5db65471a625ebd9c3c2d5f9e786977b6136f6473127089f5a1545e99"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:36:57.079472Z","signature_b64":"evScjxXGMWjzuHNisr5GdLaQQk1dllsBIT1ZqwmmtfrmtWvbskf6a83lGh/4xAG3Rs17vsob4XOg7mtKrWDKAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c03c4a9093a11072388a2d00a6910aca0c38b9f0002341d0fb0e178c06eb35a3","last_reissued_at":"2026-07-05T05:36:57.079041Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:36:57.079041Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Specializing Smaller Language Models towards Multi-Step Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Ashish Sabharwal, Hao Peng, Litu Ou, Tushar Khot, Yao Fu","submitted_at":"2023-01-30T08:51:19Z","abstract_excerpt":"The surprising ability of Large Language Models (LLMs) to perform well on complex reasoning with only few-shot chain-of-thought prompts is believed to emerge only in very large-scale models (100+ billion parameters). We show that such abilities can, in fact, be distilled down from GPT-3.5 ($\\ge$ 175B) to T5 variants ($\\le$ 11B). We propose model specialization, to specialize the model's ability towards a target task. The hypothesis is that large models (commonly viewed as larger than 100B) have strong modeling power, but are spread on a large spectrum of tasks. Small models (commonly viewed as"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2301.12726","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/2301.12726/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":"2301.12726","created_at":"2026-07-05T05:36:57.079096+00:00"},{"alias_kind":"arxiv_version","alias_value":"2301.12726v1","created_at":"2026-07-05T05:36:57.079096+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2301.12726","created_at":"2026-07-05T05:36:57.079096+00:00"},{"alias_kind":"pith_short_12","alias_value":"YA6EVEETUEIH","created_at":"2026-07-05T05:36:57.079096+00:00"},{"alias_kind":"pith_short_16","alias_value":"YA6EVEETUEIHEOEK","created_at":"2026-07-05T05:36:57.079096+00:00"},{"alias_kind":"pith_short_8","alias_value":"YA6EVEET","created_at":"2026-07-05T05:36:57.079096+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.09881","citing_title":"Toward Calibrated, Fair, and accurate Deepfake Detection","ref_index":64,"is_internal_anchor":false},{"citing_arxiv_id":"2503.08223","citing_title":"Will LLMs Scaling Hit the Wall? Breaking Barriers via Distributed Resources on Massive Edge Devices","ref_index":72,"is_internal_anchor":false},{"citing_arxiv_id":"2305.02301","citing_title":"Distilling Step-by-Step! Outperforming Larger Language Models with Less Training Data and Smaller Model Sizes","ref_index":73,"is_internal_anchor":false},{"citing_arxiv_id":"2305.18323","citing_title":"ReWOO: Decoupling Reasoning from Observations for Efficient Augmented Language Models","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2308.01825","citing_title":"Scaling Relationship on Learning Mathematical Reasoning with Large Language Models","ref_index":69,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09725","citing_title":"On-Policy Distillation with Best-of-N Teacher Rollout Selection","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09725","citing_title":"On-Policy Distillation with Best-of-N Teacher Rollout Selection","ref_index":9,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YA6EVEETUEIHEOEKFUAKNEIKZI","json":"https://pith.science/pith/YA6EVEETUEIHEOEKFUAKNEIKZI.json","graph_json":"https://pith.science/api/pith-number/YA6EVEETUEIHEOEKFUAKNEIKZI/graph.json","events_json":"https://pith.science/api/pith-number/YA6EVEETUEIHEOEKFUAKNEIKZI/events.json","paper":"https://pith.science/paper/YA6EVEET"},"agent_actions":{"view_html":"https://pith.science/pith/YA6EVEETUEIHEOEKFUAKNEIKZI","download_json":"https://pith.science/pith/YA6EVEETUEIHEOEKFUAKNEIKZI.json","view_paper":"https://pith.science/paper/YA6EVEET","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2301.12726&json=true","fetch_graph":"https://pith.science/api/pith-number/YA6EVEETUEIHEOEKFUAKNEIKZI/graph.json","fetch_events":"https://pith.science/api/pith-number/YA6EVEETUEIHEOEKFUAKNEIKZI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YA6EVEETUEIHEOEKFUAKNEIKZI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YA6EVEETUEIHEOEKFUAKNEIKZI/action/storage_attestation","attest_author":"https://pith.science/pith/YA6EVEETUEIHEOEKFUAKNEIKZI/action/author_attestation","sign_citation":"https://pith.science/pith/YA6EVEETUEIHEOEKFUAKNEIKZI/action/citation_signature","submit_replication":"https://pith.science/pith/YA6EVEETUEIHEOEKFUAKNEIKZI/action/replication_record"}},"created_at":"2026-07-05T05:36:57.079096+00:00","updated_at":"2026-07-05T05:36:57.079096+00:00"}