{"record": {"aliases": ["Reliability-Prioritized Fine-Grained Generation in Multimodal Large"], "artifacts": [{"id": "arxiv:2606.29573", "kind": "paper", "url": "https://arxiv.org/abs/2606.29573"}, {"id": "gh:weiwu2025/granfact", "kind": "repo", "url": "https://github.com/WeiWu2025/GranFact"}], "authors": ["Xiaomeng Fan", "Wei Wu", "Yuwei Wu", "Zhi Gao", "Shiyu Luo", "Mingyang Gao", "Haoyu Zhao", "Zhenxin Diao", "Yuxuan Ba", "Lijia Feng", "Yunde Jia", "Mehrtash Harandi"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2606.29573", "source_url": "https://arxiv.org/abs/2606.29573"}, "categories": ["General AI", "Multimodal"], "description": {"en": "A granularity-aware evaluation of whether multimodal models produce visually correct descriptions at reliably specific category levels."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2606.29573", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2606.29573", "title": "GranFact"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2606.29573", "modality": null, "name": "GranFact", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"admission_policy_version": "2026-09-09.1", "confidence": "Medium", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-09-25T04:40:22Z", "data_status": "primary-source-indexed", "display_eligible": "true", "origin_record_id": "bm_reliability-prioritized-fine-grained-gener_cade06a1", "origin_source": "arxiv", "recognition_confidence": "0.7", "record_sha256": "6a6c8950d4a8fb1d03d38fcb4811140b3e30717a276ee71b70767febda76e2d4", "relation": "introduces", "review_model": "codex-cli-default", "review_state": "ai-reviewed", "reviewed_at": "2026-09-24T17:21:45.202648Z", "source_url": "https://arxiv.org/abs/2606.29573"}, "publisher": null, "released": "2026-06-28", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2606.29573", "source_url": "https://arxiv.org/abs/2606.29573"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2606-29573", "source": "claire_radar", "source_benchmark_id": "2606.29573", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": ["Reliability-Prioritized Fine-Grained Generation in Multimodal Large"], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Multimodal", "attentionForecast": {"confidence": "Low", "horizon": "7d", "reason": "Its targeted reliability question and released artifacts may draw interest from multimodal evaluation researchers, though the supplied text gives limited evidence about likely reach.", "score": 61}, "capabilities": [], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-09-24T17:21:45.202648Z", "inputHash": "941c767a9a9f8c17446747d32a2db7d04e7d445db56a246c0c4bb2d20353200a", "model": "codex-cli-default"}, "curation": {"admissionPolicyVersion": "2026-09-09.1", "canonicalNameEvidence": "we develop \\textsc{GranFact}, a granularity-aware benchmark", "canonicalNameSource": "abstract", "decisionReason": "The abstract provides a formal name, an annotated evaluation task, a hierarchy-aware scoring procedure, and an explicit public code-and-data path.", "evaluationEvidence": {"reuse": "Code and data are available \\href{https://github.com/WeiWu2025/GranFact}{here}.", "scoring": "assesses both whether model predictions are visually correct and how specific the correct predictions are", "task": "expert-verified multi-object images with coarse-to-fine category annotations"}, "model": "codex-cli-default", "reviewedAt": "2026-09-24T17:21:45.202648Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "A granularity-aware evaluation of whether multimodal models produce visually correct descriptions at reliably specific category levels.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "public_reusable", "evidence": {"reasonCodes": ["benchmark term in abstract", "benchmark release stated in one sentence", "evaluation protocol evidence", "public artifact URL"], "snippet": "To investigate this problem, we develop \\textsc{GranFact}, a granularity-aware benchmark consisting of expert-verified multi-object images with coarse-to-fine category annotations."}, "familyId": "bmf_62e3782f7bef", "firstSeenAt": "2026-08-19", "heat": null, "id": "bm_reliability-prioritized-fine-grained-gener_cade06a1", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": "https://github.com/WeiWu2025/GranFact", "data": null, "paper": "https://arxiv.org/abs/2606.29573", "pdf": "https://arxiv.org/pdf/2606.29573", "project": null, "report": "https://arxiv.org/abs/2606.29573"}, "metrics": [], "motivation": "Multimodal large language models (MLLMs) are increasingly expected to generate fine-grained descriptions of visual content.", "name": "GranFact", "oneLine": "A granularity-aware evaluation of whether multimodal models produce visually correct descriptions at reliably specific category levels.", "paperTitle": "Reliability-Prioritized Fine-Grained Generation in Multimodal Large", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2606.29573", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-19T11:05:13.395391Z"}, "publications": [], "readiness": "Runnable", "recognitionConfidence": 0.7, "relation": "introduces", "releasedAt": "2026-06-28", "reviewLifecycle": {"state": "pending"}, "source": {"authors": ["Xiaomeng Fan", "Wei Wu", "Yuwei Wu", "Zhi Gao", "Shiyu Luo", "Mingyang Gao", "Haoyu Zhao", "Zhenxin Diao", "Yuxuan Ba", "Lijia Feng", "Yunde Jia", "Mehrtash Harandi"], "categories": ["cs.CV"], "id": "2606.29573", "title": "Reliability-Prioritized Fine-Grained Generation in Multimodal Large", "type": "arxiv", "url": "https://arxiv.org/abs/2606.29573"}, "sourceUpdatedAt": "2026-07-17T01:56:19Z", "topics": ["Multimodal"], "venueAttempts": [], "whyItMatters": "It evaluates correctness and useful specificity together, exposing the reliability trade-off in increasingly fine-grained visual generation."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
