{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2608.14329", "kind": "paper", "url": "https://arxiv.org/abs/2608.14329"}], "authors": ["Dipankar Sarkar"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.14329", "source_url": "https://arxiv.org/abs/2608.14329"}, "categories": ["Finance", "Language & Knowledge", "cs.CR"], "description": {"en": "Principle-Bench contains 168 cryptoasset financial-promotion scenarios mapped to two UK FCA principles, with paraphrase, adversarial keyword-stuffing, and boundary perturbations for evaluating LLM-as-judge on accuracy, paraphrase robustness, adversarial robustness, and calibration."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.14329", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.14329", "title": "Principle-Bench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2608.14329", "modality": null, "name": "Principle-Bench", "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": "unknown", "origin_record_id": "bm_principle-bench_119ba0f5", "origin_source": "arxiv", "recognition_confidence": "0.85", "record_sha256": "4741c0f9608f5f0edf03fb95e530bd9a31dc664199b07fe3099a2e0e6ebadfe8", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2608.14329"}, "publisher": null, "released": "2026-08-14", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.14329", "source_url": "https://arxiv.org/abs/2608.14329"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-14329", "source": "claire_radar", "source_benchmark_id": "2608.14329", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Mixed", "applicationDomains": ["Finance"], "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/2608.14329", "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": "Medium", "construction": "Aggregate Existing", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-20T15:22:10.044731Z", "inputHash": "7a8be2a0d792a4609acdb74afa5db195fc0626601ae2ac368b3a535b30201848", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "Principle-Bench contains 168 cryptoasset financial-promotion scenarios mapped to two UK FCA principles, with paraphrase, adversarial keyword-stuffing, and boundary perturbations for evaluating LLM-as-judge on accuracy, paraphrase robustness, adversarial robustness, and calibration.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["exact named benchmark artifact released in abstract", "benchmark release stated in one sentence"], "snippet": "We release Principle-Bench, 168 cryptoasset financial-promotion scenarios mapped to two UK FCA principles, with paraphrase, adversarial keyword-stuffing, and boundary perturbations authored under a pre-registered rubric; the first benchmark covering all four axes for principle-based regulation."}, "familyId": "bmf_dd484486344c", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_principle-bench_119ba0f5", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": ["Financial Services"], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2608.14329", "paper": "https://arxiv.org/abs/2608.14329", "pdf": "https://arxiv.org/pdf/2608.14329", "project": null, "report": "https://arxiv.org/abs/2608.14329"}, "metrics": [], "motivation": "Principle-based regulation, with evaluative standards such as \"fair, clear, and not misleading\" or \"deliver good outcomes\", cannot be reduced to binary predicates, and LLM-as-judge is increasingly used as the substitute.", "name": "Principle-Bench", "oneLine": "Principle-Bench contains 168 cryptoasset financial-promotion scenarios mapped to two UK FCA principles, with paraphrase, adversarial keyword-stuffing, and boundary perturbations for evaluating LLM-as-judge on accuracy, paraphrase robustness, adversarial robustness, and calibration.", "paperTitle": "A Four-Axis Trustworthiness Benchmark for LLM-as-Judge in Principle-Based Regulation", "primaryDomain": "Finance", "publication": {"evidence": "7 pages, 3 figures. Accepted at the KDD 2026 Workshop on Secure and Trustworthy Large Language Models (SeT-LLM), poster", "evidenceLevel": "author-claim", "evidenceUrl": "https://arxiv.org/abs/2608.14329", "source": "arxiv-comments", "status": "acceptance_claimed", "venue": "KDD 2026 Workshop on Secure and Trustworthy Large Language Models (SeT-LLM), poster", "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "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": 0.85, "relation": "introduces", "releasedAt": "2026-08-14", "source": {"authors": ["Dipankar Sarkar"], "categories": ["cs.CR", "cs.AI", "cs.CL", "cs.CY", "cs.LG"], "id": "2608.14329", "title": "A Four-Axis Trustworthiness Benchmark for LLM-as-Judge in Principle-Based Regulation", "type": "arxiv", "url": "https://arxiv.org/abs/2608.14329"}, "sourceUpdatedAt": "2026-08-14T14:20:38Z", "topics": ["cs.CR"], "venueAttempts": [{"decisionRaw": "7 pages, 3 figures. Accepted at the KDD 2026 Workshop on Secure and Trustworthy Large Language Models (SeT-LLM), poster", "evidence": [{"level": "author-claim", "observedAt": "2026-10-14T01:46:58.058199Z", "rawValue": "7 pages, 3 figures. Accepted at the KDD 2026 Workshop on Secure and Trustworthy Large Language Models (SeT-LLM), poster", "sourceType": "arxiv-comments", "sourceUrl": "https://arxiv.org/abs/2608.14329"}], "reviewStatus": "accepted", "venueName": "KDD 2026 Workshop on Secure and Trustworthy Large Language Models (SeT-LLM), poster"}], "whyItMatters": "Addresses the evaluation gap for LLM-as-judge in principle-based regulation, where standards are not binary. Provides a multi-axis assessment to inform deployment decisions, as no single method dominates all axes and adversarial inputs can significantly degrade performance."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
