List your company

Machine Translation on FLORES-200 zho_Hant (devtest)

24Tie %GRPO
22.823.42424.6May 15, 2026

Evaluation results

MethodLinks
GRPOModel Scale=600M, Experiment=A, Sample Size=50 source sentences, Evaluation Protocol=Human preference (one-sided binomial test)2026.05243.9468—0.001
Baseline NLLB-200Model Scale=600M, Decoding=beam search, Fine-tuning=none, Sample Size=50 source sentences, Evaluation Protocol=Human preference (one-sided binomial test)2026.0524——80.001