{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2608.29270", "kind": "paper", "url": "https://arxiv.org/abs/2608.29270"}], "authors": ["Hojae Han", "Jongyoon Kim", "Sanghyeok Park", "Dongwook Cheon", "Yeachan Park", "Myeong Jae Jeon", "Sunjong Choe", "Soonho Kong", "Wonseok Hur", "Seung-won Hwang", "Donghoon Hyeon"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.29270", "source_url": "https://arxiv.org/abs/2608.29270"}, "categories": ["General AI", "Language & Knowledge", "Safety"], "description": {"en": "Evaluates autoformalization systems on 178 Lean 4 problems using the SA-Pass metric, which checks compile, forward, and backward semantic alignment via shadows."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.29270", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.29270", "title": "ShadowBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2608.29270", "modality": null, "name": "ShadowBench", "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_shadowbench_349af251", "origin_source": "arxiv", "recognition_confidence": "0.5", "record_sha256": "07378e12dae9222caeeba114b428c15d34a711d7b3200a86b488fcd2b8c54c60", "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/2608.29270"}, "publisher": null, "released": "2026-08-29", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.29270", "source_url": "https://arxiv.org/abs/2608.29270"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-29270", "source": "claire_radar", "source_benchmark_id": "2608.29270", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "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": null, "hfPaperUrl": "https://huggingface.co/papers/2608.29270", "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"}, "attentionForecast": {"confidence": "Low", "horizon": "7d", "reason": "The benchmark addresses a specialized autoformalization evaluation and has challenge adoption, but lack of immediate public artifact links may slow initial traction.", "score": 62}, "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": "9f39373fe4c2792aa86ca1acaab72dbac0f12ac2cb16a439119f2780de41a28a", "model": "deepseek-v4-pro"}, "curation": {"canonicalNameEvidence": "We instantiate SA-Pass in ShadowBench, a Lean 4 full autoformalization benchmark of 178 postgraduate- to research-level problems", "canonicalNameSource": "abstract", "decisionReason": "Formally named benchmark with defined SA-Pass scoring and a mention of serving in an ICML challenge, but no public code or data links are provided in this input.", "model": "deepseek-v4-pro", "reviewedAt": "2026-09-02T05:23:11.715633Z", "state": "ai-name-audit-deferred"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "Evaluates autoformalization systems on 178 Lean 4 problems using the SA-Pass metric, which checks compile, forward, and backward semantic alignment via shadows.", "displayEligible": false, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "score_submission", "evidence": {"reasonCodes": ["coined title prefix ending in Bench or Benchmark"], "snippet": "We instantiate SA-Pass in ShadowBench, a Lean 4 full autoformalization benchmark of 178 postgraduate- to research-level problems spanning eight mathematical areas."}, "familyId": "bmf_e7957ea2baeb", "firstSeenAt": "2026-09-01", "heat": null, "id": "bm_shadowbench_349af251", "indexedAt": "2026-09-01T06:28:29.895112Z", "industrySectors": [], "links": {"code": null, "data": null, "paper": "https://arxiv.org/abs/2608.29270", "pdf": "https://arxiv.org/pdf/2608.29270", "project": null, "report": "https://arxiv.org/abs/2608.29270"}, "metrics": [], "motivation": "Autoformalization translates informal mathematical theorems into code for proof assistants such as Lean.", "name": "ShadowBench", "oneLine": "Evaluates autoformalization systems on 178 Lean 4 problems using the SA-Pass metric, which checks compile, forward, and backward semantic alignment via shadows.", "paperTitle": "SHADOWBENCH: Toward Reliable Automatic Evaluation of Semantic Alignment in Autoformalization", "primaryDomain": "General AI", "publication": {"evidence": "The 2026 Conference on Empirical Methods in Natural Language Processing", "evidenceLevel": "strong-author-metadata", "evidenceUrl": "https://arxiv.org/abs/2608.29270", "source": "arxiv-journal-reference", "status": "publication_reported", "venue": "The 2026 Conference on Empirical Methods in Natural Language Processing", "verifiedAt": "2026-09-01T06:28:29.895112Z"}, "publications": [{"evidence": [{"level": "strong-author-metadata", "observedAt": "2026-09-01T06:28:29.895112Z", "rawValue": "The 2026 Conference on Empirical Methods in Natural Language Processing", "sourceType": "arxiv-journal-reference", "sourceUrl": "https://arxiv.org/abs/2608.29270"}], "publicationStatus": "published", "venueName": "The 2026 Conference on Empirical Methods in Natural Language Processing"}], "ranking": {"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}}, "readiness": "Paper only", "recognitionConfidence": 0.5, "relation": "introduces", "releasedAt": "2026-08-29", "source": {"authors": ["Hojae Han", "Jongyoon Kim", "Sanghyuk Park", "Dongwook Cheon", "Myungjae Jeon", "Sunjong Choi", "Soonho Kong", "Wonseok Heo", "Seung-won Hwang", "Donghoon Hyeon"], "categories": ["cs.CL", "cs.AI"], "id": "2608.29270", "title": "SHADOWBENCH: Toward Reliable Automatic Evaluation of Semantic Alignment in Autoformalization", "type": "arxiv", "url": "https://arxiv.org/abs/2608.29270"}, "sourceUpdatedAt": "2026-08-29T13:49:20Z", "topics": ["Safety"], "venueAttempts": [], "whyItMatters": "Introduces a semantic alignment evaluation for autoformalization that detects type-correct but misaligned statements beyond Pass@k metrics, grounded in expert agreement."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
