{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2607.08317", "kind": "paper", "url": "https://arxiv.org/abs/2607.08317"}, {"id": "gh:matteosantelmo/reasoning-blind-spots", "kind": "repo", "url": "https://github.com/matteosantelmo/reasoning-blind-spots"}], "authors": ["Matteo Santelmo", "Xiuying Wei", "Israa Fakih", "Felix Bauer", "Juan Garcia Giraldo", "Chengkun Li", "Etienne Bamas", "Emmanuel Abbé"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2607.08317", "source_url": "https://arxiv.org/abs/2607.08317"}, "categories": ["General AI", "Multimodal"], "description": {"en": "Evaluates reasoning blind spots in language, vision-language, and image-generation models across 235 samples with structured reference solutions and taxonomy."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2607.08317", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2607.08317", "title": "Blind-Spots-Bench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["École Polytechnique Fédérale de Lausanne"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2607.08317", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2607.08317", "source_url": "https://openalex.org/W7167941704"}, "key": "claire-radar:2607.08317", "modality": null, "name": "Blind-Spots-Bench", "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_blind-spots-bench_1b74311c", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "ed217d1f4b4c023ae7de1c306371c70ef72e155429e1ed9f858c117c178eb604", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2607.08317"}, "publisher": null, "released": "2026-07-09", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2607.08317", "source_url": "https://arxiv.org/abs/2607.08317"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2607-08317", "source": "claire_radar", "source_benchmark_id": "2607.08317", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Multimodal", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": "https://github.com/matteosantelmo/reasoning-blind-spots", "githubScope": "benchmark_repo", "githubStars": 4, "hfDailySubmittedAt": "2026-07-15T00:00:00.000Z", "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": 31, "hfPaperUrl": "https://huggingface.co/papers/2607.08317", "observedAt": "2026-10-04T15:31:48.030916Z", "signalStatus": {"githubStars": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}}, "status": "fresh"}, "capabilities": [], "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": "2ee38fb7fce4dca5a602a776c08b2ff548172b2b38b3d879b249f7720491acf9", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "Evaluates reasoning blind spots in language, vision-language, and image-generation models across 235 samples with structured reference solutions and taxonomy.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["coined title prefix ending in Bench or Benchmark", "benchmark release stated in one sentence", "evaluation protocol evidence"], "snippet": "We introduce $\\texttt{blind-spots-bench}$, a benchmark designed to expose such blind spots through tasks that appear simple for humans but remain challenging for modern AI."}, "familyId": "bmf_74da405a337c", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [{"evidence": [{"basis": "existing-code-link", "linkedUrl": "https://github.com/matteosantelmo/reasoning-blind-spots", "url": "https://github.com/matteosantelmo/reasoning-blind-spots"}, {"basis": "existing-metric-association", "linkedUrl": "https://github.com/matteosantelmo/reasoning-blind-spots", "url": "https://huggingface.co/papers/2607.08317"}, {"basis": "huggingface-paper-association", "linkedUrl": "https://github.com/matteosantelmo/reasoning-blind-spots", "url": "https://huggingface.co/api/papers/2607.08317"}], "scope": "project_repo", "url": "https://github.com/matteosantelmo/reasoning-blind-spots", "verification": "source-linked"}], "status": "linked"}, "heat": null, "id": "bm_blind-spots-bench_1b74311c", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": "https://github.com/matteosantelmo/reasoning-blind-spots", "data": null, "hfPaper": "https://huggingface.co/papers/2607.08317", "paper": "https://arxiv.org/abs/2607.08317", "pdf": "https://arxiv.org/pdf/2607.08317", "project": null, "report": "https://arxiv.org/abs/2607.08317"}, "metrics": [], "motivation": "Modern AI models achieve strong performance on many established benchmarks, yet they still fail on tasks that humans find almost trivial, such as manipulating a string or drawing a dog with five legs.", "name": "Blind-Spots-Bench", "oneLine": "Evaluates reasoning blind spots in language, vision-language, and image-generation models across 235 samples with structured reference solutions and taxonomy.", "paperTitle": "Blind-Spots-Bench: Evaluating Blind Spots in Multimodal Models", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2607.08317", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "ranking": {"90d": {"confidence": "Medium", "coverage": 0.7, "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": 0.21483102985913266, "value": 4}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": 0.6123768519227606, "value": 31}}, "confidence": "Medium", "coverage": 0.7, "rank": 172, "score": 36}, "method": "attention-ranking-v13", "rank": 172, "score": 36}}, "readiness": "Runnable", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-07-09", "source": {"authors": ["Matteo Santelmo", "Xiuying Wei", "Israa Fakih", "Felix Bauer", "Juan Garcia Giraldo", "Chengkun Li", "Etienne Bamas", "Emmanuel Abb\\'e"], "categories": ["cs.AI"], "id": "2607.08317", "title": "Blind-Spots-Bench: Evaluating Blind Spots in Multimodal Models", "type": "arxiv", "url": "https://arxiv.org/abs/2607.08317"}, "sourceUpdatedAt": "2026-07-09T09:56:50Z", "topics": ["Multimodal"], "venueAttempts": [], "whyItMatters": "Serves as a diagnostic stress test exposing tasks that humans find easy but AI models struggle with, highlighting gaps not captured by existing benchmarks."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
