{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2610.15732", "kind": "paper", "url": "https://arxiv.org/abs/2610.15732"}, {"id": "gh:iriis-research/nepplural", "kind": "repo", "url": "https://github.com/iriis-research/NepPlural"}], "authors": ["Jinu Nyachhyon", "Anish Khatiwada", "Bimal Timilsina"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2610.15732", "source_url": "https://arxiv.org/abs/2610.15732"}, "categories": ["General AI", "Language & Knowledge", "cs.CL"], "description": {"en": "NepPlural is a dataset and evaluation suite for modeling societal personas in Nepali migration discourse, with human gold annotations, LLM protocol audit, encoder baselines, and human-judged steerability."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2610.15732", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2610.15732", "title": "NepPlural"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2610.15732", "modality": null, "name": "NepPlural", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"admission_policy_version": "2026-09-09.1", "confidence": "High", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-10-14T09:44:51Z", "data_status": "primary-source-indexed", "display_eligible": "true", "origin_record_id": "bm_nepplural_d39a60a6", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "6325f741039aaea3b2c51d323d37e983ec0bb18d347596b008f3295d46b9eebe", "relation": "introduces", "review_model": "deepseek-v4-pro", "review_state": "ai-reviewed", "reviewed_at": "2026-10-14T01:46:39.631755Z", "source_url": "https://arxiv.org/abs/2610.15732"}, "publisher": null, "released": "2026-10-11", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2610.15732", "source_url": "https://arxiv.org/abs/2610.15732"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2610-15732", "source": "claire_radar", "source_benchmark_id": "2610.15732", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Language & Knowledge", "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "Low-resource language benchmarks and LLM annotation audits draw moderate interdisciplinary interest but may have a narrower audience than large-scale multimodal benchmarks.", "score": 42}, "capabilities": [], "confidence": "High", "construction": "Transform Existing", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-10-14T01:46:39.631755Z", "inputHash": "fca943c885fa3d76e0f3920735e6cffa8944e3562ab1eb85bc2890930101a94b", "model": "deepseek-v4-pro"}, "curation": {"admissionPolicyVersion": "2026-09-09.1", "canonicalNameEvidence": "NepPlural: A Human-Anchored Benchmark for Persona Annotation in Nepali Migration Discourse", "canonicalNameSource": "paper_title", "decisionReason": "The benchmark provides fixed human gold labels, defined metrics, and public code/data for reproducibly evaluating persona annotation models.", "evaluationEvidence": {"reuse": "The released per-comment outputs in results/ are the fixed artifacts... The full dataset, human gold labels, and evaluation scripts are available in the repository.", "scoring": "Mean macro-F1 over the four axes on the 300 human-gold comments... Human reference: leave-one-annotator-out score (0.75)... Macro-F1 is computed over the classes present in the reference; majority-vote ties count as errors.", "task": "Given a comment text, a model must assign persona labels on four axes (Intent, Primary Driver, Value Orientation, Affect) plus an explicit No-Persona class."}, "model": "deepseek-v4-pro", "reviewedAt": "2026-10-14T01:46:39.631755Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "NepPlural is a dataset and evaluation suite for modeling societal personas in Nepali migration discourse, with human gold annotations, LLM protocol audit, encoder baselines, and human-judged steerability.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "public_reusable", "evidence": {"reasonCodes": ["exact coined title identity tied to benchmark evidence", "benchmark release stated in one sentence", "evaluation protocol evidence", "public artifact URL"], "snippet": "We introduce NepPlural, a human-anchored benchmark for modelling competing socio-cultural personas in Nepali public discourse on youth migration and brain drain."}, "familyId": "bmf_eac9ba732acd", "firstSeenAt": "2026-10-14", "heat": null, "id": "bm_nepplural_d39a60a6", "indexedAt": "2026-10-14T01:34:24.723681Z", "industrySectors": [], "links": {"code": "https://github.com/iriis-research/NepPlural", "data": null, "paper": "https://arxiv.org/abs/2610.15732", "pdf": "https://arxiv.org/pdf/2610.15732", "project": null, "report": "https://arxiv.org/abs/2610.15732"}, "metrics": [], "motivation": "Pluralistic alignment asks that language models faithfully represent a society's diverse, competing values rather than collapse them into one dominant viewpoint, yet flagship pluralistic-alignment datasets remain heavily skewed toward Western respondents.", "name": "NepPlural", "oneLine": "NepPlural is a dataset and evaluation suite for modeling societal personas in Nepali migration discourse, with human gold annotations, LLM protocol audit, encoder baselines, and human-judged steerability.", "paperTitle": "NepPlural: A Human-Anchored Benchmark for Persona Annotation in Nepali Migration Discourse", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2610.15732", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:34:24.723681Z"}, "publications": [], "publishers": [{"name": "IRIIS Research", "organizationType": "academic-lab", "role": "benchmark-publisher", "sourceUrl": "https://github.com/iriis-research/NepPlural"}], "readiness": "Runnable", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-10-11", "source": {"authors": ["Jinu Nyachhyon", "Anish Khatiwada", "Bimal Timilsina"], "categories": ["cs.CL"], "id": "2610.15732", "title": "NepPlural: A Human-Anchored Benchmark for Persona Annotation in Nepali Migration Discourse", "type": "arxiv", "url": "https://arxiv.org/abs/2610.15732"}, "sourceUpdatedAt": "2026-10-11T16:12:20Z", "topics": ["cs.CL"], "venueAttempts": [], "whyItMatters": "It offers a human-anchored resource for low-resource Nepali NLP and investigates the reliability of LLM annotations and persona control, with practical implications for pluralistic alignment."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
