Table 6 BLEU-4, ROUGE-1, ROUGE-2, ROUGE-L evaluation metrics results
From: Automatic recognition of cross-language classic entities based on large language models
Model | ChatGLM3-6B | Baichuan2-7B-Base | ||||||
|---|---|---|---|---|---|---|---|---|
Shot | Zero | One | Three | Five | Zero | One | Three | Five |
BLEU-4 | 91.54 | 91.67 | 91.79 | 91.74 | 92.80 | 92.81 | 92.75 | 92.38 |
ROUGE-1 | 97.37 | 97.38 | 97.45 | 97.41 | 97.75 | 97.78 | 97.73 | 97.62 |
ROUGE-2 | 91.55 | 91.69 | 91.86 | 91.79 | 92.75 | 92.88 | 92.81 | 92.47 |
ROUGE-L | 95.31 | 95.38 | 95.43 | 95.39 | 95.97 | 96.04 | 95.99 | 95.80 |
Model | Xunzi-Baichuan | Xunzi-GLM | ||||||
|---|---|---|---|---|---|---|---|---|
Shot | Zero | One | Three | Five | Zero | One | Three | Five |
BLEU-4 | 93.34 | 93.43 | 93.38 | 93.33 | 92.42 | 92.45 | 92.59 | 92.57 |
ROUGE-1 | 98.02 | 98.03 | 98.04 | 98.02 | 97.71 | 97.69 | 97.74 | 97.75 |
ROUGE-2 | 93.49 | 93.57 | 93.55 | 93.52 | 92.56 | 92.58 | 92.76 | 92.73 |
ROUGE-L | 96.36 | 96.41 | 96.40 | 96.39 | 95.83 | 95.83 | 95.93 | 95.93 |
