{"record": {"aliases": ["S3Gym: Can LLMs Turn Self-Testing and Self-Judging into Self-Improvement?"], "artifacts": [{"id": "arxiv:2608.31100", "kind": "paper", "url": "https://arxiv.org/abs/2608.31100"}], "authors": ["Jiajun Shi", "Siyuan Tao", "Yuhao Wu", "Zexuan Wang", "Jingyuan Zhang", "Jiaheng Liu", "Xinping Lei", "Xinrong Zhang", "Siyuan Fang", "Zhewen Tan", "Tianle Cai", "Junhao Fang", "Jiameng Huang", "Yueyang Wang", "Jinkai Liu", "Yuxuan Zhang", "Jian Yang", "Zhoujun Li", "Shen Yan", "Wenhao Huang", "Ge Zhang"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.31100", "source_url": "https://arxiv.org/abs/2608.31100"}, "categories": ["General AI", "Language & Knowledge", "Self-Evolution"], "description": {"en": "Interactive benchmark for evaluating LLM self-improvement via self-testing, self-judging, and self-improvement in text-based games."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.31100", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.31100", "title": "S3Gym"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2608.31100", "modality": null, "name": "S3Gym", "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": "false", "origin_record_id": "bm_s3gym-can-llms-turn-self-testing-and-self-_4b2932c1", "origin_source": "arxiv", "recognition_confidence": "0.6", "record_sha256": "f97010aae8b31a04d95f37ea8a6fb8ef0cdfdd24a6e771388adfa5778e3ee76e", "relation": "introduces", "review_model": "deepseek-v4-pro", "review_state": "ai-name-audit-deferred", "reviewed_at": "2026-09-02T05:23:11.715633Z", "source_url": "https://arxiv.org/abs/2608.31100"}, "publisher": null, "released": "2026-08-31", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.31100", "source_url": "https://arxiv.org/abs/2608.31100"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-31100", "source": "claire_radar", "source_benchmark_id": "2608.31100", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": ["S3Gym: Can LLMs Turn Self-Testing and Self-Judging into Self-Improvement?"], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Language & Knowledge", "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": 41, "hfPaperUrl": "https://huggingface.co/papers/2608.31100", "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", "lastSuccessfulDate": "2026-10-01", "state": "stale"}}, "status": "stale"}, "attentionForecast": {"confidence": "Low", "horizon": "7d", "reason": "Agent self-improvement is a hot topic, but the absence of a released environment limits immediate attention.", "score": 40}, "capabilities": [], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-09-02T05:23:11.715633Z", "inputHash": "70474a8cec21ab3ba3b8fcabc1e491ee79142dd4ba2ba6c903e44b2d4172df76", "model": "deepseek-v4-pro"}, "curation": {"canonicalNameEvidence": "S3Gym: Can LLMs Turn Self-Testing and Self-Judging into Self-Improvement?", "canonicalNameSource": "paper_title", "decisionReason": "No public code or environment link is provided, leaving the benchmark's availability unclear.", "model": "deepseek-v4-pro", "reviewedAt": "2026-09-02T05:23:11.715633Z", "state": "ai-name-audit-deferred"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "Interactive benchmark for evaluating LLM self-improvement via self-testing, self-judging, and self-improvement in text-based games.", "displayEligible": false, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "public_reusable", "evidence": {"reasonCodes": ["benchmark term in abstract", "benchmark release stated in one sentence", "evaluation protocol evidence"], "snippet": "We introduce \\textbf{S\\textsuperscript{3}Gym}, an interactive benchmark for evaluating LLM self-improvement through three coupled capabilities: \\textbf{Self-Testing}, \\textbf{Self-Judging}, and \\textbf{Self-Improvement}."}, "familyId": "bmf_07d1ab4c2437", "firstSeenAt": "2026-09-01", "heat": null, "id": "bm_s3gym-can-llms-turn-self-testing-and-self-_4b2932c1", "indexedAt": "2026-09-01T06:28:29.895112Z", "industrySectors": [], "links": {"code": null, "data": null, "paper": "https://arxiv.org/abs/2608.31100", "pdf": "https://arxiv.org/pdf/2608.31100", "project": null, "report": "https://arxiv.org/abs/2608.31100"}, "metrics": [], "motivation": "Large language models (LLMs) increasingly interact with external environments and accumulate substantial behavioral experience, yet existing agent benchmarks largely evaluate them as fixed policies.", "name": "S3Gym", "oneLine": "Interactive benchmark for evaluating LLM self-improvement via self-testing, self-judging, and self-improvement in text-based games.", "paperTitle": "S3Gym: Can LLMs Turn Self-Testing and Self-Judging into Self-Improvement?", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2608.31100", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-09-01T06:28:29.895112Z"}, "publications": [], "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": 0.6604260735859022, "value": 41}}, "confidence": "Low", "coverage": 0.15, "rank": null, "score": 52}, "method": "attention-ranking-v13", "rank": null, "score": 52}}, "readiness": "Paper only", "recognitionConfidence": 0.6, "relation": "introduces", "releasedAt": "2026-08-31", "source": {"authors": ["Jiajun Shi", "Siyuan Tao", "Yuhao Wu", "Zexuan Wang", "Jingyuan Zhang", "Jiaheng Liu", "Xinping Lei", "Xinrong Zhang", "Siyuan Fang", "Zhewen Tan", "Tianle Cai", "Junhao Fang", "Jiameng Huang", "Yueyang Wang", "Jinkai Liu", "Yuxuan Zhang", "Jian Yang", "Zhoujun Li", "Shen Yan", "Wenhao Huang", "Ge Zhang"], "categories": ["cs.CL"], "id": "2608.31100", "title": "S3Gym: Can LLMs Turn Self-Testing and Self-Judging into Self-Improvement?", "type": "arxiv", "url": "https://arxiv.org/abs/2608.31100"}, "sourceUpdatedAt": "2026-08-31T17:05:41Z", "topics": ["Self-Evolution"], "venueAttempts": [], "whyItMatters": "Provides a framework to diagnose how agents turn interaction experience into reliable improvements."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
