{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2607.04293", "kind": "paper", "url": "https://arxiv.org/abs/2607.04293"}, {"kind": "website", "url": "https://causalgame.github.io/"}], "authors": ["Zhenhao Chen", "Yongqiang Chen", "Chenxi Liu", "Junchi Yu", "Xiangchen Song", "Zijian Li", "Jialin Li", "Philip Torr", "Bo Han", "Kun Zhang"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2607.04293", "source_url": "https://arxiv.org/abs/2607.04293"}, "categories": ["Agents & Tool Use", "General AI", "cs.CL"], "description": {"en": "CausalGame is a benchmark for evaluating causal thinking in LLM agents through interactive games. It includes 14 scenarios with selection bias, measurement error, and hidden confounders. Agents design experimental protocols, collect data, and produce solutions with explanations, scored against analytical optima and causal-reasoning rubrics."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2607.04293", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2607.04293", "title": "CausalGame"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Carnegie Mellon University", "Hong Kong Baptist University", "University of Oxford", "New York University Abu Dhabi"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2607.04293", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2607.04293", "source_url": "https://openalex.org/W7167618004"}, "key": "claire-radar:2607.04293", "modality": null, "name": "CausalGame", "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_causalgame_db1d5cd5", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "3503b0200f0a7743e6021cf4a97b416ee78c312696d0ace6b18c7cad141d2180", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2607.04293"}, "publisher": null, "released": "2026-07-05", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2607.04293", "source_url": "https://arxiv.org/abs/2607.04293"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2607-04293", "source": "claire_radar", "source_benchmark_id": "2607.04293", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Agents & Tool Use", "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/2607.04293", "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": [], "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": "9c8fa6ded882bcecb48d786209c9442b6ea865d7ac763f993e6a4de6b33a4acb", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "CausalGame is a benchmark for evaluating causal thinking in LLM agents through interactive games. It includes 14 scenarios with selection bias, measurement error, and hidden confounders. Agents design experimental protocols, collect data, and produce solutions with explanations, scored against analytical optima and causal-reasoning rubrics.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["exact coined title identity tied to benchmark evidence", "benchmark release stated in one sentence", "public artifact URL"], "snippet": "To this end, we present CausalGame, a benchmark that evaluates the causal thinking capabilities of LLM agents through interactive games."}, "familyId": "bmf_533dd8b8578b", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [{"evidence": [{"basis": "linked-from-project", "linkedUrl": "https://github.com/CausalGame/CausalGame", "url": "https://causalgame.github.io/"}], "scope": "benchmark_repo", "url": "https://github.com/CausalGame/CausalGame", "verification": "source-linked"}, {"evidence": [{"basis": "linked-from-project", "linkedUrl": "https://github.com/viewsetting/CausalGame", "url": "https://causalgame.github.io/"}], "scope": "benchmark_repo", "url": "https://github.com/viewsetting/CausalGame", "verification": "source-linked"}], "status": "linked"}, "heat": null, "id": "bm_causalgame_db1d5cd5", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2607.04293", "paper": "https://arxiv.org/abs/2607.04293", "pdf": "https://arxiv.org/pdf/2607.04293", "project": "https://causalgame.github.io/", "report": "https://arxiv.org/abs/2607.04293"}, "metrics": [], "motivation": "Building AI Scientist agents with Large Language Models (LLMs) has recently attracted growing attention.", "name": "CausalGame", "oneLine": "CausalGame is a benchmark for evaluating causal thinking in LLM agents through interactive games. It includes 14 scenarios with selection bias, measurement error, and hidden confounders. Agents design experimental protocols, collect data, and produce solutions with explanations, scored against analytical optima and causal-reasoning rubrics.", "paperTitle": "CausalGame: Benchmarking Causal Thinking of LLM Agents in Games", "primaryDomain": "General AI", "publication": {"evidence": "Zhenhao, Yongqiang, and Chenxi contributed equally to the project. A short version is accepted at the Forty-Third International Conference on Machine Learning (ICML) 2026 as an Oral presentation. Project website https://causalgame.github.io/", "evidenceLevel": "author-claim", "evidenceUrl": "https://arxiv.org/abs/2607.04293", "source": "arxiv-comments", "status": "acceptance_claimed", "venue": "Forty-Third International Conference on Machine Learning (ICML) 2026 as an Oral presentation", "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "publishers": [{"name": "CausalGame Team", "organizationType": "academic-lab", "role": "benchmark-publisher", "sourceUrl": "https://causalgame.github.io/"}], "ranking": {}, "readiness": "Inspectable", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-07-05", "source": {"authors": ["Zhenhao Chen", "Yongqiang Chen", "Chenxi Liu", "Junchi Yu", "Xiangchen Song", "Zijian Li", "Jialin Li", "Philip Torr", "Bo Han", "Kun Zhang"], "categories": ["cs.CL", "cs.AI", "cs.LG", "stat.ML"], "id": "2607.04293", "title": "CausalGame: Benchmarking Causal Thinking of LLM Agents in Games", "type": "arxiv", "url": "https://arxiv.org/abs/2607.04293"}, "sourceUpdatedAt": "2026-07-05T13:08:51Z", "topics": ["cs.CL"], "venueAttempts": [{"decisionRaw": "Zhenhao, Yongqiang, and Chenxi contributed equally to the project. A short version is accepted at the Forty-Third International Conference on Machine Learning (ICML) 2026 as an Oral presentation. Project website https://causalgame.github.io/", "evidence": [{"level": "author-claim", "observedAt": "2026-10-14T01:46:58.058199Z", "rawValue": "Zhenhao, Yongqiang, and Chenxi contributed equally to the project. A short version is accepted at the Forty-Third International Conference on Machine Learning (ICML) 2026 as an Oral presentation. Project website https://causalgame.github.io/", "sourceType": "arxiv-comments", "sourceUrl": "https://arxiv.org/abs/2607.04293"}], "reviewStatus": "accepted", "venueName": "Forty-Third International Conference on Machine Learning (ICML) 2026 as an Oral presentation"}], "whyItMatters": "Existing AI Scientist benchmarks do not isolate causal reasoning under realistic biases. CausalGame provides a structured evaluation for this capability, offering decision value for developers of autonomous research agents seeking to assess robustness against confounders and biases."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
