Chinese Value Alignment Benchmark 1.0 (Chinese)
3,054.69Sovereignty ScoreVC-LLM
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| VC-LLMalignment_framework=adversarial alignment2026.01 | 3,054.69 | 1,434.77 | 955 | 2,064.9 | 874.83 | 836 | 4.8 | |
| Qwen2.5-7B-InstructParameters=7B, Variant=Instruct2026.01 | 3,044.68 | 1,153.83 | 884.63 | 1,563.71 | 703.89 | 733 | 4.21 | |
| Qwen2.5-72B-InstructParameters=72B, Variant=Instruct2026.01 | 3,024.65 | 1,173.9 | 924.84 | 1,764.19 | 794.39 | 766 | 4.4 | |
| Qwen2.5-14B-InstructParameters=14B, Variant=Instruct2026.01 | 2,964.55 | 1,214.03 | 884.63 | 1,633.88 | 784.33 | 746 | 4.29 | |
| Baichuan2-13B-ChatParameters=13B, Variant=Chat2026.01 | 2,624.03 | 1,063.53 | 743.89 | 1,563.71 | 663.67 | 664 | 3.82 | |
| GLM-4-9B-ChatParameters=9B, Variant=Chat2026.01 | 2,493.83 | 1,153.83 | 834.37 | 1,653.93 | 693.83 | 681 | 3.91 | |
| Baichuan2-7B-ChatParameters=7B, Variant=Chat2026.01 | 2,443.75 | 943.13 | 693.63 | 1,403.33 | 563.11 | 603 | 3.47 | |
| Yi-6B-ChatParameters=6B, Variant=Chat2026.01 | 2,423.72 | 983.27 | 703.68 | 1,323.14 | 573.17 | 599 | 3.44 | |
| GPT-4o2026.01 | 2,023.11 | 862.87 | 482.53 | 1,232.93 | 392.17 | 498 | 2.86 | |
| Ziya2-13B-ChatParameters=13B, Variant=Chat2026.01 | 2,003.08 | 812.7 | 593.11 | 1,192.83 | 422.33 | 501 | 2.88 | |
| Llama-3-Ch-8B-v3Parameters=8B, Variant=Chinese-v32026.01 | 1,812.78 | 842.8 | 512.68 | 1,112.64 | 522.89 | 479 | 2.75 | |
| TigerBot-7BParameters=7B2026.01 | 1,191.83 | 391.3 | 341.79 | 811.93 | 372.06 | 310 | 1.78 | |
| internlm2.5-7b-chatParameters=7B, Variant=Chat2026.01 | 1,061.63 | 471.57 | 291.53 | 571.36 | 372.06 | 276 | 1.59 |