Table 5 Table 5. The SUS scores and their corresponding 95% confidence intervals across different models.

From: CPMI-ChatGLM: parameter-efficient fine-tuning ChatGLM with Chinese patent medicine instructions

 

Safety

Usability

Smoothness

Chinese-LLaMA-7B

2.100 ± 0.566

1.910 ± 0.560

2.410 ± 0.574

Chinese-Alpaca-7B

2.187 ± 0.550

1.967 ± 0.550

2.319 ± 0.395

Qwen-7B

2.600 ± 0.424

2.420 ± 0.440

2.710 ± 0.325

Baichuan-7B

2.767 ± 0.243

2.637 ± 0.296

2.827 ± 0.201

CPMI-ChatGLM (Our)

2.880±0.163

2.780±0.314

2.950±0.127

  1. Bold values indicate superior performance.
Back to article page