Train models to resolve evidence boundaries in lacquerware heritage questions, from conflicting records to answers the sources cannot support.

Train models to identify people, places, events, and textual evidence accurately using Korean passages with detailed JSON annotations.

Compare festival and ritual practices using regional sources, while helping models represent local differences without overgeneralizing.

Train and evaluate Korean document understanding with TOPIK reading questions, structured answers, and supporting evidence.

Explore expert Q&A on traditional instrument roles, ensemble practice, and performance contexts, grounded in supplied sources.

Train models to give verifiable specialist answers grounded in museum collection catalogs.

The Religious Activities Chanting Audio Retrieval Dataset is a high-quality audio resource aimed at promoting intelligent recognition and analysis of language and cultural phenomena.

Explore the literal meanings, idiomatic senses, and usage conditions of German fixed expressions to improve explanations of real-world language.

Turn real-world Korean public notices into clear learner-friendly explanations with structured details about requirements and actions.

Train more reliable answers about traditional dance heritage with traceable evidence and explicit comparison of source differences.

Train models to follow event narratives, argument logic, and cross-sentence relations in formal Korean news and commentary.

Explore the pragmatic nuances and usage rules of German expressions such as doch, ja, and mal.

Train models to identify Arabic rhetorical terms, explain their definitions, and analyze their effects in context.

Train models to reason step by step about painting attributions and workshop relationships using catalog records and evidence from multiple sources.

Train models to trace painting provenance across ownership records, transfers, dates, and documentary evidence.

Train models to combine evidence across religious text passages and produce grounded, traceable answers within a defined corpus.

Connect Arabic expressions to real usage contexts, politeness strategies, and register guidance for more precise retrieval of pragmatic knowledge.

Connect real-world Korean learning questions with traceable cultural evidence to make pragmatic guidance more reliable.

Help models understand and explain real-world Korean medical instructions, with a focus on procedures, timing, and risk-related details.

Help models explain traditional theater roles, stage conventions, and performance rules using verifiable source material.