Question Answering on BIGBENCH II
100True WS ScorePromptCOS
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| PromptCOSModel=TinyLlama-chat2025.09 | 100 | 20 | 0.64 | |
| PromptCOSModel=Deepseek-d-qwen2025.09 | 100 | 24 | 0.64 | |
| PCGModel=Gemma2-it2025.09 | 100 | 22 | 0.71 | |
| PromptCOSModel=Gemma2-it2025.09 | 95 | 29 | 0.6 | |
| PRModel=TinyLlama-chat2025.09 | 83 | 75 | -0.01 | |
| PRModel=Gemma2-it2025.09 | 81 | 71 | 0.05 | |
| PRModel=Deepseek-d-qwen2025.09 | 73 | 49 | 0.13 | |
| PCGModel=TinyLlama-chat2025.09 | 67 | 54 | 0.06 | |
| PC*Model=Gemma2-it2025.09 | 65 | 2 | 0.61 | |
| PC*Model=TinyLlama-chat2025.09 | 64 | 8 | 0.52 | |
| PC*Model=Deepseek-d-qwen2025.09 | 58 | 5 | 0.51 | |
| PCGModel=Deepseek-d-qwen2025.09 | 11 | 3 | 0.02 |