ResearchBenchmarksCross-lingual Knowledge Alignment on BMLAMA-53 (test)Follow0.9062Pearson CorrelationNAVCA0.66180.725250.78870.85215Jul 20, 2025Evaluation ResultsMethodMethodLinksPearson CorrelationNAVCASentence Representatio...Sentence Representation=weighted, Model=Mistral 0.3 7B2025.070.9062MEXASentence Representatio...Sentence Representation=weighted, Model=Mistral 0.3 7B2025.070.8795NAVCASentence Representatio...Sentence Representation=last, Model=GLM 4 9B2025.070.6737NASCASentence Representatio...Sentence Representation=average, Model=Llama 3.2 3B2025.070.6712