{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2609.03588", "kind": "paper", "url": "https://arxiv.org/abs/2609.03588"}], "authors": ["Yaxing Lyu", "Shengjie Zhou", "Binbin Toh", "Pengyu Zhu", "Lijun Li"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2609.03588", "source_url": "https://arxiv.org/abs/2609.03588"}, "categories": ["Agents & Tool Use", "Factuality", "General AI", "cs.AI"], "description": {"en": "A multi-turn benchmark with 238 tasks using a user simulator, stateful tools, and environment assertions to test LLM agents' handling of knowledge conflicts across world, input, and temporal categories."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2609.03588", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2609.03588", "title": "KC-Bench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Shanghai Artificial Intelligence Laboratory", "University of Hong Kong", "Xiamen University Malaysia", "Beijing University of Posts and Telecommunications"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2609.03588", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2609.03588", "source_url": "https://openalex.org/W7208756648"}, "key": "claire-radar:2609.03588", "modality": null, "name": "KC-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": "false", "origin_record_id": "bm_kc-bench_089cdd9b", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "70eeeeb340fe753237cbb383b9bc935ee8a50f99e828ae76ac421584cebcb018", "relation": "introduces", "review_model": "deepseek-v4-pro", "review_state": "ai-name-audit-deferred", "reviewed_at": "2026-09-07T00:42:15.394147Z", "source_url": "https://arxiv.org/abs/2609.03588"}, "publisher": null, "released": "2026-09-03", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2609.03588", "source_url": "https://arxiv.org/abs/2609.03588"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2609-03588", "source": "claire_radar", "source_benchmark_id": "2609.03588", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Agents & Tool Use", "attention": {"asOf": "2026-09-04", "attemptedAt": "2026-09-04T06:36:46.783164Z", "githubRepo": null, "githubScope": null, "githubStars": null, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": null, "hfPaperUrl": "https://huggingface.co/papers/2609.03588", "observedAt": "2026-09-03T06:34:20.331209Z", "signalStatus": {"githubStars": {"state": "not_applicable"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"checkedAt": "2026-09-04T06:36:46.783164Z", "state": "unavailable"}}, "status": "unavailable"}, "attentionForecast": {"confidence": "Low", "horizon": "7d", "reason": "Without public artifacts or a clear release path, there is insufficient basis to forecast attention beyond the paper itself.", "score": 0}, "capabilities": ["Factuality"], "confidence": "High", "construction": "Interactive Environment", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-09-07T00:42:15.394147Z", "inputHash": "3ce1f97a0b94bc4663b61f9b9a7f3a8e1608001971f4af86d8dd28e535c77626", "model": "deepseek-v4-pro"}, "curation": {"canonicalNameEvidence": "We introduce KC-Bench, a controlled multi-turn benchmark for measuring this capability across world-knowledge conflicts, input inconsistencies, and multi-source temporal conflicts.", "canonicalNameSource": "abstract", "decisionReason": "The abstract mentions evaluation infrastructure but no artifact links or data release, so it cannot be confirmed as a publicly reusable benchmark.", "model": "deepseek-v4-pro", "reviewedAt": "2026-09-07T00:42:15.394147Z", "state": "ai-name-audit-deferred"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "A multi-turn benchmark with 238 tasks using a user simulator, stateful tools, and environment assertions to test LLM agents' handling of knowledge conflicts across world, input, and temporal categories.", "displayEligible": false, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "score_submission", "evidence": {"reasonCodes": ["coined title prefix ending in Bench or Benchmark", "benchmark release stated in one sentence", "evaluation protocol evidence"], "snippet": "We introduce KC-Bench, a controlled multi-turn benchmark for measuring this capability across world-knowledge conflicts, input inconsistencies, and multi-source temporal conflicts."}, "familyId": "bmf_30f2a9cedf7f", "firstSeenAt": "2026-09-04", "heat": null, "id": "bm_kc-bench_089cdd9b", "indexedAt": "2026-09-04T00:41:56.430646Z", "industrySectors": [], "links": {"code": null, "data": null, "paper": "https://arxiv.org/abs/2609.03588", "pdf": "https://arxiv.org/pdf/2609.03588", "project": null, "report": "https://arxiv.org/abs/2609.03588"}, "metrics": [], "motivation": "As LLMs increasingly act through tools, they must reconcile user instructions, parametric knowledge, and dynamic environmental observations before taking actions.", "name": "KC-Bench", "oneLine": "A multi-turn benchmark with 238 tasks using a user simulator, stateful tools, and environment assertions to test LLM agents' handling of knowledge conflicts across world, input, and temporal categories.", "paperTitle": "KC-Bench: A Dynamic Interactive Benchmark for Evaluating Knowledge Conflicts in LLM Agents", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2609.03588", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-09-04T00:41:56.430646Z"}, "publications": [], "ranking": {"30d": {"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": 30}, "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}, "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}, "today": {"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": 1, "historyDate": "2026-09-03", "rank": null, "score": null, "windowDays": 1}, "level": {"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}, "llmAttentionForecast": {"maxContribution": 0.25, "percentile": 0.0, "role": "bonus", "value": 0}}, "confidence": "Low", "coverage": 0.0, "rank": null, "score": 0}, "method": "attention-ranking-v13", "rank": null, "score": 0}}, "readiness": "Paper only", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-09-03", "source": {"authors": ["Yaxing Lyu", "Shengjie Zhou", "Binbin Toh", "Pengyu Zhu", "Lijun Li"], "categories": ["cs.AI"], "id": "2609.03588", "title": "KC-Bench: A Dynamic Interactive Benchmark for Evaluating Knowledge Conflicts in LLM Agents", "type": "arxiv", "url": "https://arxiv.org/abs/2609.03588"}, "sourceUpdatedAt": "2026-09-03T09:35:14Z", "topics": ["cs.AI"], "venueAttempts": [], "whyItMatters": "It isolates conflict-handling behavior in agents, aiming to improve action safety when user instructions, model knowledge, and environmental observations diverge."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
