{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:K5GYBH5QVJ562GCB2GEU7MAUT6","short_pith_number":"pith:K5GYBH5Q","schema_version":"1.0","canonical_sha256":"574d809fb0aa7bed1841d1894fb0149fa1dcd6f68dd03c30a51b1aea5503ed10","source":{"kind":"arxiv","id":"2310.11716","version":1},"attestation_state":"computed","paper":{"title":"Reflection-Tuning: Data Recycling Improves LLM Instruction-Tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Heng Huang, Jiuhai Chen, Jiuxiang Gu, Lichang Chen, Ming Li, Shwai He, Tianyi Zhou","submitted_at":"2023-10-18T05:13:47Z","abstract_excerpt":"Recent advancements in Large Language Models (LLMs) have expanded the horizons of natural language understanding and generation. Notably, the output control and alignment with the input of LLMs can be refined through instruction tuning. However, as highlighted in several studies, low-quality data in the training set are usually detrimental to instruction tuning, resulting in inconsistent or even misleading LLM outputs. We propose a novel method, termed \"reflection-tuning,\" which addresses the problem by self-improvement and judging capabilities of LLMs. This approach utilizes an oracle LLM to "},"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":"2310.11716","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-10-18T05:13:47Z","cross_cats_sorted":[],"title_canon_sha256":"a3f3a7b8678d1fa65e1ac62664dbcfc476ca9b688badaa71be19b47f2bbd78f6","abstract_canon_sha256":"511b9d97c77f78f983694b8604fbcd69ea687ef0bf15e47a597ad8fe0d8a2401"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:02:11.309827Z","signature_b64":"LzL3xp1lu6X2TiWpphYd9lyjXyulsdU3umwCCnHsjjhmz1mNBldTbpS67QH1bJzg3qUBBiBA4w8oOZWcirdbCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"574d809fb0aa7bed1841d1894fb0149fa1dcd6f68dd03c30a51b1aea5503ed10","last_reissued_at":"2026-07-05T07:02:11.309451Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:02:11.309451Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Reflection-Tuning: Data Recycling Improves LLM Instruction-Tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Heng Huang, Jiuhai Chen, Jiuxiang Gu, Lichang Chen, Ming Li, Shwai He, Tianyi Zhou","submitted_at":"2023-10-18T05:13:47Z","abstract_excerpt":"Recent advancements in Large Language Models (LLMs) have expanded the horizons of natural language understanding and generation. Notably, the output control and alignment with the input of LLMs can be refined through instruction tuning. However, as highlighted in several studies, low-quality data in the training set are usually detrimental to instruction tuning, resulting in inconsistent or even misleading LLM outputs. We propose a novel method, termed \"reflection-tuning,\" which addresses the problem by self-improvement and judging capabilities of LLMs. This approach utilizes an oracle LLM to "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.11716","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/2310.11716/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":"2310.11716","created_at":"2026-07-05T07:02:11.309507+00:00"},{"alias_kind":"arxiv_version","alias_value":"2310.11716v1","created_at":"2026-07-05T07:02:11.309507+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.11716","created_at":"2026-07-05T07:02:11.309507+00:00"},{"alias_kind":"pith_short_12","alias_value":"K5GYBH5QVJ56","created_at":"2026-07-05T07:02:11.309507+00:00"},{"alias_kind":"pith_short_16","alias_value":"K5GYBH5QVJ562GCB","created_at":"2026-07-05T07:02:11.309507+00:00"},{"alias_kind":"pith_short_8","alias_value":"K5GYBH5Q","created_at":"2026-07-05T07:02:11.309507+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2402.06922","citing_title":"Whispers in the Machine: Confidentiality in Agentic Systems","ref_index":45,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/K5GYBH5QVJ562GCB2GEU7MAUT6","json":"https://pith.science/pith/K5GYBH5QVJ562GCB2GEU7MAUT6.json","graph_json":"https://pith.science/api/pith-number/K5GYBH5QVJ562GCB2GEU7MAUT6/graph.json","events_json":"https://pith.science/api/pith-number/K5GYBH5QVJ562GCB2GEU7MAUT6/events.json","paper":"https://pith.science/paper/K5GYBH5Q"},"agent_actions":{"view_html":"https://pith.science/pith/K5GYBH5QVJ562GCB2GEU7MAUT6","download_json":"https://pith.science/pith/K5GYBH5QVJ562GCB2GEU7MAUT6.json","view_paper":"https://pith.science/paper/K5GYBH5Q","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2310.11716&json=true","fetch_graph":"https://pith.science/api/pith-number/K5GYBH5QVJ562GCB2GEU7MAUT6/graph.json","fetch_events":"https://pith.science/api/pith-number/K5GYBH5QVJ562GCB2GEU7MAUT6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/K5GYBH5QVJ562GCB2GEU7MAUT6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/K5GYBH5QVJ562GCB2GEU7MAUT6/action/storage_attestation","attest_author":"https://pith.science/pith/K5GYBH5QVJ562GCB2GEU7MAUT6/action/author_attestation","sign_citation":"https://pith.science/pith/K5GYBH5QVJ562GCB2GEU7MAUT6/action/citation_signature","submit_replication":"https://pith.science/pith/K5GYBH5QVJ562GCB2GEU7MAUT6/action/replication_record"}},"created_at":"2026-07-05T07:02:11.309507+00:00","updated_at":"2026-07-05T07:02:11.309507+00:00"}