{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2608.03505", "kind": "paper", "url": "https://arxiv.org/abs/2608.03505"}], "authors": ["Jinhong Jeong", "Seungyeop Yi", "Sangah Lee", "Youngjae Yu"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.03505", "source_url": "https://arxiv.org/abs/2608.03505"}, "categories": ["Factuality", "General AI", "Language & Knowledge", "cs.CL"], "description": {"en": "ConlangBench evaluates large language models on translation and vocabulary learning across 21 constructed languages, with a corpus of over 21 million conlang-English parallel sentence pairs and 321K vocabulary entries. The benchmark includes bidirectional translation tasks and learning-curve analysis for models trained on eight conlangs with sufficient parallel data."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.03505", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.03505", "title": "ConlangBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Yonsei University", "Seoul National University"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2608.03505", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2608.03505", "source_url": "https://openalex.org/W7172488844"}, "key": "claire-radar:2608.03505", "modality": null, "name": "ConlangBench", "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_conlangbench_b9af66c3", "origin_source": "arxiv", "recognition_confidence": "0.85", "record_sha256": "39b2728b2f2a114316e64c5523196e036a3c8e061112fa1c48f1fbb5adb7a2c3", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2608.03505"}, "publisher": null, "released": "2026-08-04", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.03505", "source_url": "https://arxiv.org/abs/2608.03505"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-03505", "source": "claire_radar", "source_benchmark_id": "2608.03505", "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/2608.03505", "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": ["Factuality"], "confidence": "Medium", "construction": "Transform Existing", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-20T15:22:10.044731Z", "inputHash": "cc778933dc9f0238630f3713d79065c4088d322a404a6dd081accf04595e492c", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "ConlangBench evaluates large language models on translation and vocabulary learning across 21 constructed languages, with a corpus of over 21 million conlang-English parallel sentence pairs and 321K vocabulary entries. The benchmark includes bidirectional translation tasks and learning-curve analysis for models trained on eight conlangs with sufficient parallel data.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["coined title prefix ending in Bench or Benchmark", "benchmark release stated in one sentence"], "snippet": "We present ConlangBench, the first large-scale benchmark for evaluating and training LLMs on 21 existing conlangs."}, "familyId": "bmf_fd59187772ea", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_conlangbench_b9af66c3", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2608.03505", "paper": "https://arxiv.org/abs/2608.03505", "pdf": "https://arxiv.org/pdf/2608.03505", "project": null, "report": "https://arxiv.org/abs/2608.03505"}, "metrics": [], "motivation": "Constructed languages (conlangs) are intentionally created human languages with a rich tradition of linguistic creativity.", "name": "ConlangBench", "oneLine": "ConlangBench evaluates large language models on translation and vocabulary learning across 21 constructed languages, with a corpus of over 21 million conlang-English parallel sentence pairs and 321K vocabulary entries. The benchmark includes bidirectional translation tasks and learning-curve analysis for models trained on eight conlangs with sufficient parallel data.", "paperTitle": "ConlangBench: Exploring Language Knowledge and Learning in LLMs through Diverse Constructed Languages", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2608.03505", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "publishers": [{"name": "ConlangBench Team", "organizationType": "academic-lab", "role": "benchmark-publisher", "sourceUrl": "https://arxiv.org/abs/2608.03505"}], "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": "2026-08-04", "source": {"authors": ["Jinhong Jeong", "Seungyeop Yi", "Sangah Lee", "Youngjae Yu"], "categories": ["cs.CL"], "id": "2608.03505", "title": "ConlangBench: Exploring Language Knowledge and Learning in LLMs through Diverse Constructed Languages", "type": "arxiv", "url": "https://arxiv.org/abs/2608.03505"}, "sourceUpdatedAt": "2026-08-05T01:59:09Z", "topics": ["cs.CL"], "venueAttempts": [], "whyItMatters": "ConlangBench addresses the gap in evaluating LLMs on low-resource languages with diverse linguistic structures. It provides a controlled testbed for studying language acquisition and cross-lingual transfer, offering practical insights for model developers targeting underrepresented languages."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
