Chinese Subjective Alignment on CompassArena
31.4Win RateInternLM2-Chat-20B
Evaluation Results
| Method | Links | |
|---|---|---|
| InternLM2-Chat-20BCategory=~20B Models, Parameter Size=20B, Strategy=RLHF2024.03 | 31.4 | |
| InternLM2-Chat-7BCategory=~7B Models, Parameter Size=7B, Strategy=RLHF2024.03 | 28.7 | |
| Qwen-14B-ChatCategory=~20B Models, Parameter Size=14B2024.03 | 24.8 | |
| GPT-3.5Category=API Models2024.03 | 24.7 | |
| Baichuan2-13B-ChatCategory=~20B Models, Parameter Size=13B2024.03 | 20.5 | |
| Mixtral-8x7B-Instruct-v0.1Category=~20B Models, Parameter Size=8x7B2024.03 | 18.9 | |
| ChatGLM3-6BCategory=~7B Models, Parameter Size=6B2024.03 | 17.2 | |
| Baichuan2-7B-ChatCategory=~7B Models, Parameter Size=7B2024.03 | 16.1 | |
| Qwen-7B-ChatCategory=~7B Models, Parameter Size=7B2024.03 | 15.8 | |
| InternLM2-Chat-20B-SFTCategory=~20B Models, Parameter Size=20B, Strategy=SFT2024.03 | 15.2 | |
| Mistral-7B-Instruct-v0.2Category=~7B Models, Parameter Size=7B2024.03 | 14.2 | |
| InternLM2-Chat-7B-SFTCategory=~7B Models, Parameter Size=7B, Strategy=SFT2024.03 | 14.1 | |
| Llama2-13B-ChatCategory=~20B Models, Parameter Size=13B2024.03 | 9.8 | |
| Llama2-7B-ChatCategory=~7B Models, Parameter Size=7B2024.03 | 6.5 |