{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:PJOWPOFFWLEMB7JYLTWCWOGEE5","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"24b193d28ef5af944ab35cb2be4e90913f09b547ee2d6b7a86d57d3933323322","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-11-15T18:58:31Z","title_canon_sha256":"bc7d3a69bb86e42ea12f690bae4d1046c5a3e7378c8f824482aa58f70d6e11b9"},"schema_version":"1.0","source":{"id":"2411.10440","kind":"arxiv","version":6}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2411.10440","created_at":"2026-05-17T23:38:48Z"},{"alias_kind":"arxiv_version","alias_value":"2411.10440v6","created_at":"2026-05-17T23:38:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.10440","created_at":"2026-05-17T23:38:48Z"},{"alias_kind":"pith_short_12","alias_value":"PJOWPOFFWLEM","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"PJOWPOFFWLEMB7JY","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"PJOWPOFF","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:13e915080d642f564690e6a83bece4f74456784c64d3f8bce016e8c9814efb47","target":"graph","created_at":"2026-05-17T23:38:48Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"with only 100k training samples and test-time scaling, LLaVA-CoT not only outperforms its base model by 9.4% on a wide range of multimodal reasoning benchmarks, but also surpasses the performance of larger and even closed-source models, such as Gemini-1.5-pro, GPT-4o-mini, and Llama-3.2-90B-Vision-Instruct."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the human-provided structured reasoning annotations in the LLaVA-CoT-100k dataset faithfully capture effective multistage reasoning without introducing systematic biases or annotation artifacts that the model simply memorizes."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"LLaVA-CoT adds autonomous multistage reasoning to vision-language models, delivering 9.4% gains over its base model and outperforming larger models like Gemini-1.5-pro on reasoning benchmarks via a 100k annotated dataset and SWIRES test-time scaling."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"By training on structured four-stage annotations, LLaVA-CoT lets vision-language models reason autonomously and outperform larger models with only 100k samples."}],"snapshot_sha256":"e493f6aa8471655f9d546fbcf7c8330b5f777cb99e5eb857e8c488adfc19b644"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"d6d3ef0647160794211088bf4ccb2eea9985534d9d19b104dbc1bb6fa3ca30ca"},"paper":{"abstract_excerpt":"Large language models have demonstrated substantial advancements in reasoning capabilities. However, current Vision-Language Models (VLMs) often struggle to perform systematic and structured reasoning, especially when handling complex visual question-answering tasks. In this work, we introduce LLaVA-CoT, a large VLM designed to conduct autonomous multistage reasoning. Unlike chain-of-thought prompting, LLaVA-CoT independently engages in sequential stages of summarization, visual interpretation, logical reasoning, and conclusion generation. This structured approach enables LLaVA-CoT to achieve ","authors_text":"Guowei Xu, Hao Li, Lichao Sun, Li Yuan, Peng Jin, Yibing Song, Ziang Wu","cross_cats":[],"headline":"By training on structured four-stage annotations, LLaVA-CoT lets vision-language models reason autonomously and outperform larger models with only 100k samples.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-11-15T18:58:31Z","title":"LLaVA-CoT: Let Vision Language Models Reason Step-by-Step"},"references":{"count":68,"internal_anchors":3,"resolved_work":68,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"https : / / opencompass","work_id":"fbaa7c31-1483-4ab8-b7d9-4fbb67f9e1d5","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Available at: https://www","work_id":"82bd6267-6caa-4ddf-a95a-11d7c2fe5afc","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Gpt-4o system card, 2024","work_id":"adccb5da-15dc-4b4e-af45-0758f6342e7d","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Variational best-of-n alignment, 2024","work_id":"975889b4-94a1-4f87-808c-751080915571","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Neuro-symbolic visual reasoning: Disentangling","work_id":"55532825-9fd4-4d62-aacd-bf7d8915c455","year":2020}],"snapshot_sha256":"253a79e483a70865339e1d8a1c2569b0a9850de6bf157d4eebdf0d057f8dcf63"},"source":{"id":"2411.10440","kind":"arxiv","version":6},"verdict":{"created_at":"2026-05-16T11:31:07.126334Z","id":"7c3f801f-376d-4a69-b65a-a4a169c90d26","model_set":{"reader":"grok-4.3"},"one_line_summary":"LLaVA-CoT adds autonomous multistage reasoning to vision-language models, delivering 9.4% gains over its base model and outperforming larger models like Gemini-1.5-pro on reasoning benchmarks via a 100k annotated dataset and SWIRES test-time scaling.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"By training on structured four-stage annotations, LLaVA-CoT lets vision-language models reason autonomously and outperform larger models with only 100k samples.","strongest_claim":"with only 100k training samples and test-time scaling, LLaVA-CoT not only outperforms its base model by 9.4% on a wide range of multimodal reasoning benchmarks, but also surpasses the performance of larger and even closed-source models, such as Gemini-1.5-pro, GPT-4o-mini, and Llama-3.2-90B-Vision-Instruct.","weakest_assumption":"That the human-provided structured reasoning annotations in the LLaVA-CoT-100k dataset faithfully capture effective multistage reasoning without introducing systematic biases or annotation artifacts that the model simply memorizes."}},"verdict_id":"7c3f801f-376d-4a69-b65a-a4a169c90d26"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:9564e47b1d9cfef5e6485778671107d3aeb87c28483786b1c1438bd60d3a9fd8","target":"record","created_at":"2026-05-17T23:38:48Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"24b193d28ef5af944ab35cb2be4e90913f09b547ee2d6b7a86d57d3933323322","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-11-15T18:58:31Z","title_canon_sha256":"bc7d3a69bb86e42ea12f690bae4d1046c5a3e7378c8f824482aa58f70d6e11b9"},"schema_version":"1.0","source":{"id":"2411.10440","kind":"arxiv","version":6}},"canonical_sha256":"7a5d67b8a5b2c8c0fd385cec2b38c4275c7481a89b5dba265e12bb5c41fff2e1","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"7a5d67b8a5b2c8c0fd385cec2b38c4275c7481a89b5dba265e12bb5c41fff2e1","first_computed_at":"2026-05-17T23:38:48.018188Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:38:48.018188Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"UkgNV9bTsupHpKs+26qKxT0bokw9xmVDBb3m942XxORKkhcDfgpoepNhXHVDbbOWJxPPCoHJeo1N4A64CbsQCg==","signature_status":"signed_v1","signed_at":"2026-05-17T23:38:48.018791Z","signed_message":"canonical_sha256_bytes"},"source_id":"2411.10440","source_kind":"arxiv","source_version":6}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:9564e47b1d9cfef5e6485778671107d3aeb87c28483786b1c1438bd60d3a9fd8","sha256:13e915080d642f564690e6a83bece4f74456784c64d3f8bce016e8c9814efb47"],"state_sha256":"30db60ff2cf843f558d4b305232ba134c92fbd8e67f6a791cd873a4c5dfbeaa0"}