{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:76KZSPXYRUL3Q65Y2ZZ5BL3OD5","short_pith_number":"pith:76KZSPXY","schema_version":"1.0","canonical_sha256":"ff95993ef88d17b87bb8d673d0af6e1f614dc44c96f2c405a6ce0213540c0b9f","source":{"kind":"arxiv","id":"2503.01754","version":3},"attestation_state":"computed","paper":{"title":"SDRT: Enhance Vision-Language Models by Self-Distillation with Diverse Reasoning Traces","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Guande Wu, Huan Song, Lin Lee Cheong, Panpan Xu, Qiaojing Yan, Yawei Wang, Yijun Tian","submitted_at":"2025-03-03T17:24:42Z","abstract_excerpt":"Reasoning is increasingly crucial for various tasks. While chain-of-thought prompting enables large language models to leverage reasoning effectively, harnessing the reasoning capabilities of Vision-Language Models (VLMs) remains challenging. To solve this problem, we propose a novel self-distillation framework that enhances the reasoning capabilities of the model. The proposed framework introduces several key innovations. We start by employing a prompt library tailored to visual reasoning tasks to generate diverse in-context questions and utilize a two-step reasoning procedure to derive reaso"},"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":"2503.01754","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-03-03T17:24:42Z","cross_cats_sorted":[],"title_canon_sha256":"35f9a93700726acae95b6e9b752a7ee46983701473394c34d10df170d24eb016","abstract_canon_sha256":"ebe25bd5279f9dc7035950a494a4272f8399fa8ed7dd6c07a19cd6a81361dc2b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:35:47.173149Z","signature_b64":"u2JwVwTRS1wGmUoWJ49AEMSeraXovBsX5eMaGkNODfjEBnGxJPfCG/U/IATlS1RBUECW2WZSyQUjQ2m9ek0NDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ff95993ef88d17b87bb8d673d0af6e1f614dc44c96f2c405a6ce0213540c0b9f","last_reissued_at":"2026-07-05T10:35:47.172588Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:35:47.172588Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SDRT: Enhance Vision-Language Models by Self-Distillation with Diverse Reasoning Traces","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Guande Wu, Huan Song, Lin Lee Cheong, Panpan Xu, Qiaojing Yan, Yawei Wang, Yijun Tian","submitted_at":"2025-03-03T17:24:42Z","abstract_excerpt":"Reasoning is increasingly crucial for various tasks. While chain-of-thought prompting enables large language models to leverage reasoning effectively, harnessing the reasoning capabilities of Vision-Language Models (VLMs) remains challenging. To solve this problem, we propose a novel self-distillation framework that enhances the reasoning capabilities of the model. The proposed framework introduces several key innovations. We start by employing a prompt library tailored to visual reasoning tasks to generate diverse in-context questions and utilize a two-step reasoning procedure to derive reaso"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.01754","kind":"arxiv","version":3},"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/2503.01754/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":"2503.01754","created_at":"2026-07-05T10:35:47.172648+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.01754v3","created_at":"2026-07-05T10:35:47.172648+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.01754","created_at":"2026-07-05T10:35:47.172648+00:00"},{"alias_kind":"pith_short_12","alias_value":"76KZSPXYRUL3","created_at":"2026-07-05T10:35:47.172648+00:00"},{"alias_kind":"pith_short_16","alias_value":"76KZSPXYRUL3Q65Y","created_at":"2026-07-05T10:35:47.172648+00:00"},{"alias_kind":"pith_short_8","alias_value":"76KZSPXY","created_at":"2026-07-05T10:35:47.172648+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.11651","citing_title":"Hide to See: Reasoning-prefix Masking for Visual-anchored Thinking in VLM Distillation","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11651","citing_title":"Hide to See: Reasoning-prefix Masking for Visual-anchored Thinking in VLM Distillation","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12034","citing_title":"Boosting Omni-Modal Language Models: Staged Post-Training with Visually Debiased Evaluation","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11651","citing_title":"Hide to See: Reasoning-prefix Masking for Visual-anchored Thinking in VLM Distillation","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12034","citing_title":"Boosting Omni-Modal Language Models: Staged Post-Training with Visually Debiased Evaluation","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11651","citing_title":"Hide to See: Reasoning-prefix Masking for Visual-anchored Thinking in VLM Distillation","ref_index":31,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/76KZSPXYRUL3Q65Y2ZZ5BL3OD5","json":"https://pith.science/pith/76KZSPXYRUL3Q65Y2ZZ5BL3OD5.json","graph_json":"https://pith.science/api/pith-number/76KZSPXYRUL3Q65Y2ZZ5BL3OD5/graph.json","events_json":"https://pith.science/api/pith-number/76KZSPXYRUL3Q65Y2ZZ5BL3OD5/events.json","paper":"https://pith.science/paper/76KZSPXY"},"agent_actions":{"view_html":"https://pith.science/pith/76KZSPXYRUL3Q65Y2ZZ5BL3OD5","download_json":"https://pith.science/pith/76KZSPXYRUL3Q65Y2ZZ5BL3OD5.json","view_paper":"https://pith.science/paper/76KZSPXY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.01754&json=true","fetch_graph":"https://pith.science/api/pith-number/76KZSPXYRUL3Q65Y2ZZ5BL3OD5/graph.json","fetch_events":"https://pith.science/api/pith-number/76KZSPXYRUL3Q65Y2ZZ5BL3OD5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/76KZSPXYRUL3Q65Y2ZZ5BL3OD5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/76KZSPXYRUL3Q65Y2ZZ5BL3OD5/action/storage_attestation","attest_author":"https://pith.science/pith/76KZSPXYRUL3Q65Y2ZZ5BL3OD5/action/author_attestation","sign_citation":"https://pith.science/pith/76KZSPXYRUL3Q65Y2ZZ5BL3OD5/action/citation_signature","submit_replication":"https://pith.science/pith/76KZSPXYRUL3Q65Y2ZZ5BL3OD5/action/replication_record"}},"created_at":"2026-07-05T10:35:47.172648+00:00","updated_at":"2026-07-05T10:35:47.172648+00:00"}