{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:TLBSRFGXBUFB72INGDAQLJSHQO","short_pith_number":"pith:TLBSRFGX","schema_version":"1.0","canonical_sha256":"9ac32894d70d0a1fe90d30c105a64783b9aee80ff78f356b518c54fff0f409fc","source":{"kind":"arxiv","id":"2505.16475","version":1},"attestation_state":"computed","paper":{"title":"ReflectEvo: Improving Meta Introspection of Small LLMs by Learning Self-Reflection","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Jiaqi Li, Quansen Wang, Songchun Zhu, Xiaobo Wang, Xinyi Dong, Yang Liu, Zhizhuo Yang, Zilong Zheng, Zixia Jia","submitted_at":"2025-05-22T10:03:05Z","abstract_excerpt":"We present a novel pipeline, ReflectEvo, to demonstrate that small language models (SLMs) can enhance meta introspection through reflection learning. This process iteratively generates self-reflection for self-training, fostering a continuous and self-evolving process. Leveraging this pipeline, we construct ReflectEvo-460k, a large-scale, comprehensive, self-generated reflection dataset with broadened instructions and diverse multi-domain tasks. Building upon this dataset, we demonstrate the effectiveness of reflection learning to improve SLMs' reasoning abilities using SFT and DPO with remark"},"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":"2505.16475","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2025-05-22T10:03:05Z","cross_cats_sorted":[],"title_canon_sha256":"9b9079fa2f472cec13d2c59c0a8e68f98b5ba72c1c72e9c07d2646d7a972119a","abstract_canon_sha256":"965091e29c0568c6a2092428ce6cb1f182b5bb8fb69034e6950247fdaf35d6a7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:07:36.125387Z","signature_b64":"/hrbPyDvrY2m8sPRAINydpcEOqsFluk/JTM6wmMMnB38mpzdRZHLPoGvTLTDPL6z/TAMcaCdJda89F9YVH1UCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9ac32894d70d0a1fe90d30c105a64783b9aee80ff78f356b518c54fff0f409fc","last_reissued_at":"2026-07-05T11:07:36.124802Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:07:36.124802Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ReflectEvo: Improving Meta Introspection of Small LLMs by Learning Self-Reflection","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Jiaqi Li, Quansen Wang, Songchun Zhu, Xiaobo Wang, Xinyi Dong, Yang Liu, Zhizhuo Yang, Zilong Zheng, Zixia Jia","submitted_at":"2025-05-22T10:03:05Z","abstract_excerpt":"We present a novel pipeline, ReflectEvo, to demonstrate that small language models (SLMs) can enhance meta introspection through reflection learning. This process iteratively generates self-reflection for self-training, fostering a continuous and self-evolving process. Leveraging this pipeline, we construct ReflectEvo-460k, a large-scale, comprehensive, self-generated reflection dataset with broadened instructions and diverse multi-domain tasks. Building upon this dataset, we demonstrate the effectiveness of reflection learning to improve SLMs' reasoning abilities using SFT and DPO with remark"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.16475","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/2505.16475/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":"2505.16475","created_at":"2026-07-05T11:07:36.124862+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.16475v1","created_at":"2026-07-05T11:07:36.124862+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.16475","created_at":"2026-07-05T11:07:36.124862+00:00"},{"alias_kind":"pith_short_12","alias_value":"TLBSRFGXBUFB","created_at":"2026-07-05T11:07:36.124862+00:00"},{"alias_kind":"pith_short_16","alias_value":"TLBSRFGXBUFB72IN","created_at":"2026-07-05T11:07:36.124862+00:00"},{"alias_kind":"pith_short_8","alias_value":"TLBSRFGX","created_at":"2026-07-05T11:07:36.124862+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.20189","citing_title":"SOLAR: A Self-Optimizing Open-Ended Autonomous Agent for Lifelong Learning and Continual Adaptation","ref_index":54,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TLBSRFGXBUFB72INGDAQLJSHQO","json":"https://pith.science/pith/TLBSRFGXBUFB72INGDAQLJSHQO.json","graph_json":"https://pith.science/api/pith-number/TLBSRFGXBUFB72INGDAQLJSHQO/graph.json","events_json":"https://pith.science/api/pith-number/TLBSRFGXBUFB72INGDAQLJSHQO/events.json","paper":"https://pith.science/paper/TLBSRFGX"},"agent_actions":{"view_html":"https://pith.science/pith/TLBSRFGXBUFB72INGDAQLJSHQO","download_json":"https://pith.science/pith/TLBSRFGXBUFB72INGDAQLJSHQO.json","view_paper":"https://pith.science/paper/TLBSRFGX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.16475&json=true","fetch_graph":"https://pith.science/api/pith-number/TLBSRFGXBUFB72INGDAQLJSHQO/graph.json","fetch_events":"https://pith.science/api/pith-number/TLBSRFGXBUFB72INGDAQLJSHQO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TLBSRFGXBUFB72INGDAQLJSHQO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TLBSRFGXBUFB72INGDAQLJSHQO/action/storage_attestation","attest_author":"https://pith.science/pith/TLBSRFGXBUFB72INGDAQLJSHQO/action/author_attestation","sign_citation":"https://pith.science/pith/TLBSRFGXBUFB72INGDAQLJSHQO/action/citation_signature","submit_replication":"https://pith.science/pith/TLBSRFGXBUFB72INGDAQLJSHQO/action/replication_record"}},"created_at":"2026-07-05T11:07:36.124862+00:00","updated_at":"2026-07-05T11:07:36.124862+00:00"}