Crossing Linguistic Horizons

Finetuning and Comprehensive Evaluation of Vietnamese Large Language Models

Zero-Shot Toxicity Detection Leaderboard

Models UiT-ViCTSD UiT-ViHSD
AC F1 AR ECE A@10 AC F1 AR ECE A@10
URA-LLaMa 70B 0.61 ± 0.01 0.52 ± 0.01 0.77 ± 0.01 0.17 ± 0.01 0.97 ± 0.01 0.38 ± 0.01 0.34 ± 0.01 0.74 ± 0.01 0.25 ± 0.01 0.91 ± 0.01
URA-LLaMa 13B 0.46 ± 0.01 0.28 ± 0.03 0.53 ± 0.02 0.22 ± 0.01 0.48 ± 0.03 0.33 ± 0.01 0.18 ± 0.00 0.60 ± 0.01 0.35 ± 0.01 0.54 ± 0.02
URA-LLaMa 7B 0.25 ± 0.01 0.19 ± 0.01 0.53 ± 0.01 0.38 ± 0.01 0.13 ± 0.02 0.19 ± 0.00 0.13 ± 0.00 0.55 ± 0.01 0.46 ± 0.01 0.13 ± 0.01
LLaMa-2 13B 0.16 ± 0.01 0.14 ± 0.00 0.40 ± 0.01 0.50 ± 0.01 0.24 ± 0.02 0.09 ± 0.00 0.13 ± 0.00 0.38 ± 0.01 0.63 ± 0.00 0.10 ± 0.01
LLaMa-2 7B 0.13 ± 0.01 0.14 ± 0.01 0.45 ± 0.02 0.69 ± 0.01 0.09 ± 0.01 0.03 ± 0.00 0.05 ± 0.01 0.56 ± 0.01 0.75 ± 0.00 0.00 ± 0.00
Vietcuna 7B 0.09 ± 0.00 0.07 ± 0.00 0.50 ± 0.00 0.41 ± 0.00 0.10 ± 0.03 0.07 ± 0.00 0.04 ± 0.00 0.50 ± 0.00 0.26 ± 0.00 0.07 ± 0.01
GPT-3.5 0.75 ± 0.01 0.61 ± 0.02 - 0.25 ± 0.01 0.80 ± 0.04 0.55 ± 0.01 0.42 ± 0.01 - 0.22 ± 0.01 0.55 ± 0.02
GPT-4 0.89 ± 0.01 0.69 ± 0.01 - 0.39 ± 0.01 0.89 ± 0.03 0.75 ± 0.01 0.53 ± 0.01 - 0.42 ± 0.01 0.75 ± 0.02