{"record": {"aliases": ["SAKE: Software Architectural Knowledge Evaluation Benchmark for Large Language Models"], "artifacts": [{"id": "arxiv:2606.29520", "kind": "paper", "url": "https://arxiv.org/abs/2606.29520"}], "authors": ["Tiziano Santilli", "Francesco Daghero", "Mayhar Tourchi Moghaddam"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2606.29520", "source_url": "https://arxiv.org/abs/2606.29520"}, "categories": ["Code & Software", "Factuality", "General AI", "cs.SE"], "description": {"en": "SAKE is a standardized multiple-choice evaluation of software-architecture knowledge in language models, covering eight architectural categories and four context-length levels."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2606.29520", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2606.29520", "title": "SAKE"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["University of Southern Denmark"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2606.29520", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2606.29520", "source_url": "https://openalex.org/W7166730336"}, "key": "claire-radar:2606.29520", "modality": null, "name": "SAKE", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"admission_policy_version": "2026-09-09.1", "confidence": "Medium", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-09-25T04:40:22Z", "data_status": "primary-source-indexed", "display_eligible": "true", "origin_record_id": "bm_sake-software-architectural-knowledge-eval_e9685637", "origin_source": "arxiv", "recognition_confidence": "0.6", "record_sha256": "f0fc42ca453796000fdf636423d20aeb3d0b5a89f8e11fb3890b1bab5712dd1e", "relation": "introduces", "review_model": "codex-cli-default", "review_state": "ai-reviewed", "reviewed_at": "2026-09-24T18:50:12.626375Z", "source_url": "https://arxiv.org/abs/2606.29520"}, "publisher": null, "released": "2026-06-28", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2606.29520", "source_url": "https://arxiv.org/abs/2606.29520"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2606-29520", "source": "claire_radar", "source_benchmark_id": "2606.29520", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": ["SAKE: Software Architectural Knowledge Evaluation Benchmark for Large Language Models"], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Code & Software", "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "Its specialized software-engineering focus and evaluation across proprietary and open-weight models may attract moderate near-term research attention.", "score": 4}, "capabilities": ["Factuality"], "confidence": "Medium", "construction": "Transform Existing", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-09-24T18:50:12.626375Z", "inputHash": "f69e9cfc20abdc47daaec0b2f7dac2654bf3849347ec627988b80074347b6fe6", "model": "codex-cli-default"}, "curation": {"admissionPolicyVersion": "2026-09-09.1", "canonicalNameEvidence": "SAKE (Software Architectural Knowledge Evaluation)", "canonicalNameSource": "abstract", "decisionReason": "The supplied abstract provides a formal name, a defined task, accuracy-based comparison, a standardized and reproducible protocol, and an explicit open-source release of the benchmark and evaluation scripts.", "evaluationEvidence": {"reuse": "SAKE, its evaluation scripts, and all results are released as open source", "scoring": "Overall accuracy is high", "task": "assessing software architectural knowledge in LLMs"}, "model": "codex-cli-default", "reviewedAt": "2026-09-24T18:50:12.626375Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "SAKE is a standardized multiple-choice evaluation of software-architecture knowledge in language models, covering eight architectural categories and four context-length levels.", "displayEligible": true, "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 SAKE (Software Architectural Knowledge Evaluation), a standardized and reproducible benchmark for assessing software architectural knowledge in LLMs."}, "familyId": "bmf_15093c71123e", "firstSeenAt": "2026-08-19", "heat": null, "id": "bm_sake-software-architectural-knowledge-eval_e9685637", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": null, "paper": "https://arxiv.org/abs/2606.29520", "pdf": "https://arxiv.org/pdf/2606.29520", "project": null, "report": "https://arxiv.org/abs/2606.29520"}, "metrics": [], "motivation": "Large Language Models (LLMs) are increasingly used as assistants across the software development lifecycle, yet their ability to reason about software architecture remains largely unmeasured.", "name": "SAKE", "oneLine": "SAKE is a standardized multiple-choice evaluation of software-architecture knowledge in language models, covering eight architectural categories and four context-length levels.", "paperTitle": "SAKE: Software Architectural Knowledge Evaluation Benchmark for Large Language Models", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2606.29520", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-19T11:05:13.395391Z"}, "publications": [], "readiness": "Paper only", "recognitionConfidence": 0.6, "relation": "introduces", "releasedAt": "2026-06-28", "reviewLifecycle": {"state": "pending"}, "source": {"authors": ["Tiziano Santilli", "Francesco Daghero", "Mayhar Tourchi Moghaddam"], "categories": ["cs.SE", "cs.AI", "cs.DB"], "id": "2606.29520", "title": "SAKE: Software Architectural Knowledge Evaluation Benchmark for Large Language Models", "type": "arxiv", "url": "https://arxiv.org/abs/2606.29520"}, "sourceUpdatedAt": "2026-06-28T17:31:14Z", "topics": ["cs.SE"], "venueAttempts": [], "whyItMatters": "It provides a repeatable way to track architectural reasoning capabilities that are not exercised by syntactic or algorithmic benchmarks."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
