{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2605.25510", "kind": "paper", "url": "https://arxiv.org/abs/2605.25510"}, {"id": "gh:michigannlp/kidbench", "kind": "repo", "url": "https://github.com/MichiganNLP/kidbench"}], "authors": ["Samee Arif", "Angana Borah", "Rada Mihalcea"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2605.25510", "source_url": "https://arxiv.org/abs/2605.25510"}, "categories": ["General AI", "Safety", "Safety & Trustworthiness"], "description": {"en": "KIDBench evaluates child-facing safety of LLMs for ages 7-11 using realistic queries and multi-turn child-actor simulations, scored by an LLM-as-a-Judge rubric."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2605.25510", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2605.25510", "title": "KIDBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2605.25510", "modality": null, "name": "KIDBench", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"confidence": "Medium", "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_kidbench_e196ca1f", "origin_source": "arxiv", "recognition_confidence": "0.85", "record_sha256": "848c2d53af1863767b0b28194e968266e769f50d684d30a67e183bdbcd394539", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2605.25510"}, "publisher": null, "released": "2026-05-25", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2605.25510", "source_url": "https://arxiv.org/abs/2605.25510"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2605-25510", "source": "claire_radar", "source_benchmark_id": "2605.25510", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Safety & Trustworthiness", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": "https://github.com/MichiganNLP/kidbench", "githubScope": "benchmark_repo", "githubStars": 1, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": null, "hfPaperUrl": "https://huggingface.co/papers/2605.25510", "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": {"checkedAt": "2026-10-04T15:31:48.030916Z", "state": "unavailable"}}, "status": "fresh"}, "capabilities": [], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-20T15:22:10.044731Z", "inputHash": "b9f62aa683b7a3015a20ce3331eeaac39dc91d4faa87f4b9ea92a774f67e38e7", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "KIDBench evaluates child-facing safety of LLMs for ages 7-11 using realistic queries and multi-turn child-actor simulations, scored by an LLM-as-a-Judge rubric.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["exact named benchmark artifact released in abstract", "benchmark release stated in one sentence"], "snippet": "We introduce KIDBench, a benchmark for evaluating child-facing LLM safety for ages 7-11 using a developmental-psychology-grounded LLM-as-a-Judge rubric."}, "familyId": "bmf_d5c09a3ec477", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [{"evidence": [{"basis": "linked-from-report", "linkedUrl": "https://github.com/MichiganNLP/kidbench", "url": "https://arxiv.org/abs/2605.25510"}, {"basis": "linked-from-report", "linkedUrl": "https://github.com/MichiganNLP/kidbench.", "url": "https://arxiv.org/abs/2605.25510"}], "scope": "benchmark_repo", "url": "https://github.com/MichiganNLP/kidbench", "verification": "source-linked"}], "status": "linked"}, "heat": null, "id": "bm_kidbench_e196ca1f", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": "https://github.com/MichiganNLP/kidbench", "data": null, "hfPaper": "https://huggingface.co/papers/2605.25510", "paper": "https://arxiv.org/abs/2605.25510", "pdf": "https://arxiv.org/pdf/2605.25510", "project": null, "report": "https://arxiv.org/abs/2605.25510"}, "metrics": [], "motivation": "Children increasingly have access to Large Language Models (LLMs), which may expose them to responses that are developmentally inappropriate or require age-sensitive safety, guidance, and boundaries.", "name": "KIDBench", "oneLine": "KIDBench evaluates child-facing safety of LLMs for ages 7-11 using realistic queries and multi-turn child-actor simulations, scored by an LLM-as-a-Judge rubric.", "paperTitle": "The Age of Curiosity Meets the Age of AI: Benchmarking Child Safety in Large Language Models", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2605.25510", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "ranking": {}, "readiness": "Runnable", "recognitionConfidence": 0.85, "relation": "introduces", "releasedAt": "2026-05-25", "source": {"authors": ["Samee Arif and Angana Borah and Rada Mihalcea"], "categories": ["cs.CL"], "id": "2605.25510", "title": "The Age of Curiosity Meets the Age of AI: Benchmarking Child Safety in Large Language Models", "type": "arxiv", "url": "https://arxiv.org/abs/2605.25510"}, "sourceUpdatedAt": "2026-05-26T09:53:31Z", "topics": ["Safety"], "venueAttempts": [], "whyItMatters": "It addresses a gap in LLM safety evaluation by focusing on age-appropriate responses for children, with a novel rubric and multi-turn assessment."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
