Crossing Linguistic Horizons

Finetuning and Comprehensive Evaluation of Vietnamese Large Language Models

Robustness-Aware Knowledge Leaderboard

Models ZaloE2E ViMMRC
EM F1 AC F1 AR ECE A@10
URA-LLaMa 70B 0.23 ± 0.00 0.37 ± 0.00 0.65 ± 0.00 0.53 ± 0.00 0.84 ± 0.00 0.11 ± 0.00 0.77 ± 0.00
URA-LLaMa 13B 0.18 ± 0.00 0.30 ± 0.00 0.41 ± 0.00 0.34 ± 0.00 0.61 ± 0.00 0.22 ± 0.00 0.58 ± 0.00
URA-LLaMa 7B 0.10 ± 0.00 0.18 ± 0.00 0.33 ± 0.02 0.28 ± 0.02 0.61 ± 0.01 0.19 ± 0.02 0.33 ± 0.06
LLaMa-2 13B 0.13 ± 0.00 0.21 ± 0.00 0.39 ± 0.00 0.31 ± 0.00 0.56 ± 0.00 0.46 ± 0.00 0.33 ± 0.00
LLaMa-2 7B 0.02 ± 0.00 0.05 ± 0.00 0.26 ± 0.01 0.20 ± 0.01 0.51 ± 0.01 0.46 ± 0.01 0.13 ± 0.03
Vietcuna 7B 0.05 ± 0.00 0.15 ± 0.00 0.26 ± 0.01 0.14 ± 0.00 0.50 ± 0.00 0.01 ± 0.01 0.21 ± 0.07
MixSUra 8x7B 0.13 ± - 0.24 ± - 0.57 ± - 0.45 ± - 0.53 ± - 0.35 ± - 0.58 ± -
GPT-3.5 0.45 ± 0.01 0.61 ± 0.01 0.90 ± 0.01 0.72 ± 0.04 - 0.65 ± 0.01 0.88 ± 0.07
GPT-4 0.44 ± 0.01 0.61 ± 0.01 0.91 ± 0.01 0.73 ± 0.07 - 0.66 ± 0.07 0.88 ± 0.04