{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:FB5JQX5UAE5FSVJFWHE5Q4BOBC","short_pith_number":"pith:FB5JQX5U","schema_version":"1.0","canonical_sha256":"287a985fb4013a595525b1c9d8702e08bd93485ef7d137fb5d9312946b40422b","source":{"kind":"arxiv","id":"2305.02301","version":2},"attestation_state":"computed","paper":{"title":"Distilling Step-by-Step! Outperforming Larger Language Models with Less Training Data and Smaller Model Sizes","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Alexander Ratner, Cheng-Yu Hsieh, Chen-Yu Lee, Chih-Kuan Yeh, Chun-Liang Li, Hootan Nakhost, Ranjay Krishna, Tomas Pfister, Yasuhisa Fujii","submitted_at":"2023-05-03T17:50:56Z","abstract_excerpt":"Deploying large language models (LLMs) is challenging because they are memory inefficient and compute-intensive for practical applications. In reaction, researchers train smaller task-specific models by either finetuning with human labels or distilling using LLM-generated labels. However, finetuning and distillation require large amounts of training data to achieve comparable performance to LLMs. We introduce Distilling step-by-step, a new mechanism that (a) trains smaller models that outperform LLMs, and (b) achieves so by leveraging less training data needed by finetuning or distillation. Ou"},"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":"2305.02301","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-03T17:50:56Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"5da5918948ac3443d91f6343d40495382da60ee1dde34469622d771d6a1923a8","abstract_canon_sha256":"dee095e9fa75bea685fa1fa46dff2ac9fc3407a9c72bf447c543c2a137a782ea"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T20:43:58.375927Z","signature_b64":"qr0VDtHStr8rByyVxRat5oF3LQHN9yU5TGSFJqvRkkp9pwFzer1EzUVmPZat2v59Zp9L+371Bi4B3MPEvzySCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"287a985fb4013a595525b1c9d8702e08bd93485ef7d137fb5d9312946b40422b","last_reissued_at":"2026-05-21T20:43:58.373761Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T20:43:58.373761Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Distilling Step-by-Step! Outperforming Larger Language Models with Less Training Data and Smaller Model Sizes","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Alexander Ratner, Cheng-Yu Hsieh, Chen-Yu Lee, Chih-Kuan Yeh, Chun-Liang Li, Hootan Nakhost, Ranjay Krishna, Tomas Pfister, Yasuhisa Fujii","submitted_at":"2023-05-03T17:50:56Z","abstract_excerpt":"Deploying large language models (LLMs) is challenging because they are memory inefficient and compute-intensive for practical applications. In reaction, researchers train smaller task-specific models by either finetuning with human labels or distilling using LLM-generated labels. However, finetuning and distillation require large amounts of training data to achieve comparable performance to LLMs. We introduce Distilling step-by-step, a new mechanism that (a) trains smaller models that outperform LLMs, and (b) achieves so by leveraging less training data needed by finetuning or distillation. Ou"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.02301","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/2305.02301/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":"2305.02301","created_at":"2026-05-21T20:43:58.374133+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.02301v2","created_at":"2026-05-21T20:43:58.374133+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.02301","created_at":"2026-05-21T20:43:58.374133+00:00"},{"alias_kind":"pith_short_12","alias_value":"FB5JQX5UAE5F","created_at":"2026-05-21T20:43:58.374133+00:00"},{"alias_kind":"pith_short_16","alias_value":"FB5JQX5UAE5FSVJF","created_at":"2026-05-21T20:43:58.374133+00:00"},{"alias_kind":"pith_short_8","alias_value":"FB5JQX5U","created_at":"2026-05-21T20:43:58.374133+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":23,"internal_anchor_count":23,"sample":[{"citing_arxiv_id":"2505.16120","citing_title":"LLM-Powered AI Agent Systems and Their Applications in Industry","ref_index":91,"is_internal_anchor":true},{"citing_arxiv_id":"2510.26745","citing_title":"Deep sequence models tend to memorize geometrically; it is unclear why","ref_index":74,"is_internal_anchor":true},{"citing_arxiv_id":"2605.10933","citing_title":"DECO: Sparse Mixture-of-Experts with Dense-Comparable Performance on End-Side Devices","ref_index":69,"is_internal_anchor":true},{"citing_arxiv_id":"2605.17774","citing_title":"Internalizing Tool Knowledge in Small Language Models via QLoRA Fine-Tuning","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2605.19394","citing_title":"EmbGen: Teaching with Reassembled Corpora","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2502.10248","citing_title":"Step-Video-T2V Technical Report: The Practice, Challenges, and Future of Video Foundation Model","ref_index":179,"is_internal_anchor":true},{"citing_arxiv_id":"2507.21954","citing_title":"Fine-Tuning Code Language Models to Detect Cross-Language Bugs","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2305.15717","citing_title":"The False Promise of Imitating Proprietary LLMs","ref_index":294,"is_internal_anchor":true},{"citing_arxiv_id":"2510.18117","citing_title":"Online In-Context Distillation for Low-Resource Vision Language Models","ref_index":11,"is_internal_anchor":true},{"citing_arxiv_id":"2602.11157","citing_title":"Response-Based Knowledge Distillation for Multilingual Jailbreak Prevention Unwittingly Compromises Safety","ref_index":20,"is_internal_anchor":true},{"citing_arxiv_id":"2401.05459","citing_title":"Personal LLM Agents: Insights and Survey about the Capability, Efficiency and Security","ref_index":258,"is_internal_anchor":true},{"citing_arxiv_id":"2601.00823","citing_title":"Energy-Aware Routing to Large Reasoning Models","ref_index":17,"is_internal_anchor":true},{"citing_arxiv_id":"2404.14294","citing_title":"A Survey on Efficient Inference for Large Language Models","ref_index":120,"is_internal_anchor":true},{"citing_arxiv_id":"2604.02371","citing_title":"Internalized Reasoning for Long-Context Visual Document Understanding","ref_index":19,"is_internal_anchor":true},{"citing_arxiv_id":"2605.10933","citing_title":"DECO: Sparse Mixture-of-Experts with Dense-Comparable Performance on End-Side Devices","ref_index":69,"is_internal_anchor":true},{"citing_arxiv_id":"2605.11290","citing_title":"ReAD: Reinforcement-Guided Capability Distillation for Large Language Models","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08094","citing_title":"MedThink: Enhancing Diagnostic Accuracy in Small Models via Teacher-Guided Reasoning Correction","ref_index":2,"is_internal_anchor":true},{"citing_arxiv_id":"2605.10933","citing_title":"DECO: Sparse Mixture-of-Experts with Dense-Comparable Performance on End-Side Devices","ref_index":69,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08838","citing_title":"Generating Leakage-Free Benchmarks for Robust RAG Evaluation","ref_index":6,"is_internal_anchor":true},{"citing_arxiv_id":"2604.24380","citing_title":"Structural Pruning of Large Vision Language Models: A Comprehensive Study on Pruning Dynamics, Recovery, and Data Efficiency","ref_index":37,"is_internal_anchor":true},{"citing_arxiv_id":"2604.18936","citing_title":"Fine-Tuning Small Reasoning Models for Quantum Field Theory","ref_index":40,"is_internal_anchor":true},{"citing_arxiv_id":"2605.07783","citing_title":"Chain-based Distillation for Effective Initialization of Variable-Sized Small Language Models","ref_index":84,"is_internal_anchor":true},{"citing_arxiv_id":"2605.02035","citing_title":"A Multimodal Dataset for Visually Grounded Ambiguity in Machine Translation","ref_index":40,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FB5JQX5UAE5FSVJFWHE5Q4BOBC","json":"https://pith.science/pith/FB5JQX5UAE5FSVJFWHE5Q4BOBC.json","graph_json":"https://pith.science/api/pith-number/FB5JQX5UAE5FSVJFWHE5Q4BOBC/graph.json","events_json":"https://pith.science/api/pith-number/FB5JQX5UAE5FSVJFWHE5Q4BOBC/events.json","paper":"https://pith.science/paper/FB5JQX5U"},"agent_actions":{"view_html":"https://pith.science/pith/FB5JQX5UAE5FSVJFWHE5Q4BOBC","download_json":"https://pith.science/pith/FB5JQX5UAE5FSVJFWHE5Q4BOBC.json","view_paper":"https://pith.science/paper/FB5JQX5U","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.02301&json=true","fetch_graph":"https://pith.science/api/pith-number/FB5JQX5UAE5FSVJFWHE5Q4BOBC/graph.json","fetch_events":"https://pith.science/api/pith-number/FB5JQX5UAE5FSVJFWHE5Q4BOBC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FB5JQX5UAE5FSVJFWHE5Q4BOBC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FB5JQX5UAE5FSVJFWHE5Q4BOBC/action/storage_attestation","attest_author":"https://pith.science/pith/FB5JQX5UAE5FSVJFWHE5Q4BOBC/action/author_attestation","sign_citation":"https://pith.science/pith/FB5JQX5UAE5FSVJFWHE5Q4BOBC/action/citation_signature","submit_replication":"https://pith.science/pith/FB5JQX5UAE5FSVJFWHE5Q4BOBC/action/replication_record"}},"created_at":"2026-05-21T20:43:58.374133+00:00","updated_at":"2026-05-21T20:43:58.374133+00:00"}