{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2607.14543", "kind": "paper", "url": "https://arxiv.org/abs/2607.14543"}], "authors": ["Huaigang Yang", "Ya Li", "Min Ren", "Bo Dai", "Zhenliang Zhang", "Zhaofeng He"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2607.14543", "source_url": "https://arxiv.org/abs/2607.14543"}, "categories": ["Robotics & Embodied AI", "Safety"], "description": {"en": "SafeRelBench evaluates VLM-driven embodied agents on household tasks with process-level safety constraints, focusing on spatial relations such as support, containment, and proximity. It includes 507 executable samples (248 spatial-relation, 259 control) and measures whether agents satisfy safety conditions before risk-prone actions, alongside task success."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2607.14543", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2607.14543", "title": "SafeRelBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Beijing University of Posts and Telecommunications", "Beijing Institute for General Artificial Intelligence"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2607.14543", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2607.14543", "source_url": "https://openalex.org/W7169591555"}, "key": "claire-radar:2607.14543", "modality": null, "name": "SafeRelBench", "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_saferelbench_483807a4", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "014faa4f5c01b49608cd1d441aec8b6469d14502c40d5f4b00feaa99d6ce6490", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2607.14543"}, "publisher": null, "released": "2026-07-16", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2607.14543", "source_url": "https://arxiv.org/abs/2607.14543"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2607-14543", "source": "claire_radar", "source_benchmark_id": "2607.14543", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["Robotics & Embodied AI"], "area": "Robotics & Embodied AI", "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": 0, "hfPaperUrl": "https://huggingface.co/papers/2607.14543", "observedAt": "2026-10-04T15:31:48.030916Z", "signalStatus": {"githubStars": {"state": "not_applicable"}, "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": "90b9a8458215916985e3beec8e3cda0da201989b6161f7a78a6812b7884e99e6", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "SafeRelBench evaluates VLM-driven embodied agents on household tasks with process-level safety constraints, focusing on spatial relations such as support, containment, and proximity. It includes 507 executable samples (248 spatial-relation, 259 control) and measures whether agents satisfy safety conditions before risk-prone actions, alongside task success.", "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": "To address this gap, we introduce SAFERELBENCH, a spatial-relation-aware safety benchmark with 507 executable evaluation samples, including 248 spatial-relation samples and 259 non-spatial control samples."}, "familyId": "bmf_e7af30443a9d", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_saferelbench_483807a4", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": ["Robotics"], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2607.14543", "paper": "https://arxiv.org/abs/2607.14543", "pdf": "https://arxiv.org/pdf/2607.14543", "project": null, "report": "https://arxiv.org/abs/2607.14543"}, "metrics": [], "motivation": "Vision-language models (VLMs) are increasingly used as the reasoning backbone of embodied agents, enabling robots to interpret visual scenes, follow language instructions, and plan multi-step actions.", "name": "SafeRelBench", "oneLine": "SafeRelBench evaluates VLM-driven embodied agents on household tasks with process-level safety constraints, focusing on spatial relations such as support, containment, and proximity. It includes 507 executable samples (248 spatial-relation, 259 control) and measures whether agents satisfy safety conditions before risk-prone actions, alongside task success.", "paperTitle": "SafeRelBench: A Spatial-Relation-Aware Benchmark for Process-Level Safety in VLM-Driven Embodied Agents", "primaryDomain": "Robotics & Embodied AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2607.14543", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "publishers": [{"name": "Authors of SafeRelBench", "organizationType": "academic-lab", "role": "benchmark-publisher", "sourceUrl": "https://arxiv.org/abs/2607.14543"}], "ranking": {"90d": {"confidence": "Low", "coverage": 0.15, "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": 0.0, "value": 0}}, "confidence": "Low", "coverage": 0.15, "rank": null, "score": 43}, "method": "attention-ranking-v13", "rank": null, "score": 43}}, "readiness": "Paper only", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-07-16", "source": {"authors": ["Huaigang Yang", "Ya Li", "Min Ren", "Bo Dai", "Zhenliang Zhang", "Zhaofeng He"], "categories": ["cs.RO", "cs.AI"], "id": "2607.14543", "title": "SafeRelBench: A Spatial-Relation-Aware Benchmark for Process-Level Safety in VLM-Driven Embodied Agents", "type": "arxiv", "url": "https://arxiv.org/abs/2607.14543"}, "sourceUpdatedAt": "2026-07-16T03:59:44Z", "topics": ["Safety"], "venueAttempts": [], "whyItMatters": "Safety in embodied agents depends on spatial awareness during action sequences, not just final outcomes. SafeRelBench fills a gap by quantifying process-level safety compliance, enabling comparison across agents and highlighting the need for improved spatial reasoning in safe planning."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
