{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2608.13267", "kind": "paper", "url": "https://arxiv.org/abs/2608.13267"}, {"kind": "website", "url": "https://scifigbench.nlp4sci.com"}], "authors": ["Paul Osemudiame Oamen", "Owusu-Banahene Osei", "Ananya Mukherjee", "Christian Greisinger", "Steffen Eger", "Pius Onobhayedo", "Wei Zhao"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.13267", "source_url": "https://arxiv.org/abs/2608.13267"}, "categories": ["General AI", "Language & Knowledge", "Reasoning"], "description": {"en": "SciFigBench is a diagnostic VLM benchmark for scientific figure understanding, covering perception, reasoning, and behavioral reliability under uncertainty."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.13267", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.13267", "title": "SciFigBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["University of Aberdeen", "International Institute of Information Technology, Hyderabad", "University of Technology Nuremberg", "University of Southern California"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2608.13267", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2608.13267", "source_url": "https://openalex.org/W7203499167"}, "key": "claire-radar:2608.13267", "modality": null, "name": "SciFigBench", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"confidence": "High", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-10-14T09:44:51Z", "data_status": "primary-source-indexed", "display_eligible": "unknown", "origin_record_id": "bm_scifigbench_9f0ee596", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "f200ea994367357f10a61423ad3a683639ba311429f83164b6157e02c97be6d3", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2608.13267"}, "publisher": null, "released": "2026-08-13", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.13267", "source_url": "https://arxiv.org/abs/2608.13267"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-13267", "source": "claire_radar", "source_benchmark_id": "2608.13267", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Language & Knowledge", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": null, "githubScope": null, "githubStars": null, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": null, "hfPaperUrl": "https://huggingface.co/papers/2608.13267", "observedAt": "2026-10-04T03:02:42.754940Z", "signalStatus": {"githubStars": {"state": "not_applicable"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "state": "unavailable"}}, "status": "unavailable"}, "capabilities": ["Reasoning"], "confidence": "High", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-20T15:22:10.044731Z", "inputHash": "ac9e5125c28750f209758db7177029cd0e9805d066722bcf9f5cec57385dc57a", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "SciFigBench is a diagnostic VLM benchmark for scientific figure understanding, covering perception, reasoning, and behavioral reliability under uncertainty.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["exact named benchmark artifact released in abstract", "benchmark release stated in one sentence", "evaluation protocol evidence", "public artifact URL"], "snippet": "We introduce SciFigBench, a diagnostic VLM benchmark for scientific figure understanding that jointly evaluates perception, reasoning, and behavioral reliability under uncertainty."}, "familyId": "bmf_aefd11b3865b", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_scifigbench_9f0ee596", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2608.13267", "paper": "https://arxiv.org/abs/2608.13267", "pdf": "https://arxiv.org/pdf/2608.13267", "project": "https://scifigbench.nlp4sci.com", "report": "https://arxiv.org/abs/2608.13267"}, "metrics": [], "motivation": "Existing vision-language model (VLM) benchmarks emphasize perception and reasoning accuracy (how well VLMs describe and reason about what they see in an image), with limited attention to behavioral reliability under uncertainty (how they behave when visual evidence is missing or misleading).", "name": "SciFigBench", "oneLine": "SciFigBench is a diagnostic VLM benchmark for scientific figure understanding, covering perception, reasoning, and behavioral reliability under uncertainty.", "paperTitle": "How Do VLMs Behave When Blind or Misled? Behavioral Evaluation of VLMs on Scientific Figures", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2608.13267", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "ranking": {"90d": {"confidence": "Low", "coverage": 0.0, "growth": {"components": {"githubStars": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfDatasetDownloads": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfPaperUpvotes": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "elapsedDays": null, "historyDate": null, "rank": null, "score": null, "windowDays": 90}, "level": {"components": {"githubStars": {"normalization": "log-max", "normalized": null, "value": null}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "rank": null, "score": null}, "method": "attention-ranking-v13", "rank": null, "score": null}}, "readiness": "Inspectable", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-08-13", "source": {"authors": ["Paul Osemudiame Oamen", "Owusu-Banahene Osei", "Ananya Mukherjee", "Christian Greisinger", "Steffen Eger", "Pius Onobhayedo", "and Wei Zhao"], "categories": ["cs.CL", "cs.AI", "cs.CV", "cs.LG"], "id": "2608.13267", "title": "How Do VLMs Behave When Blind or Misled? Behavioral Evaluation of VLMs on Scientific Figures", "type": "arxiv", "url": "https://arxiv.org/abs/2608.13267"}, "sourceUpdatedAt": "2026-08-13T14:06:35Z", "topics": ["Reasoning"], "venueAttempts": [], "whyItMatters": "It addresses the gap in evaluating VLMs' behavior when visual evidence is missing or misleading, critical for scientific workflows."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
