LaoBench is the first large-scale multidimensional benchmark for Lao, revealing that even strong multilingual LLMs lag behind human experts in culturally grounded reasoning and translation fidelity.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
LaoBench: A Large-Scale Multidimensional Lao Benchmark for Large Language Models
LaoBench is the first large-scale multidimensional benchmark for Lao, revealing that even strong multilingual LLMs lag behind human experts in culturally grounded reasoning and translation fidelity.