Quantization Detection on Math
0.532AUCRUT
Evaluation results
| Method | Links | |
|---|---|---|
| RUTModel=Llama, Quantization=4-bit2025.06 | 0.532 | |
| MMDModel=Llama, Quantization=4-bit2025.06 | 0.468 | |
| KSModel=Llama, Quantization=4-bit2025.06 | 0.412 | |
| RUTModel=Mistral, Quantization=4-bit2025.06 | 0.36 | |
| RUTModel=Gemma, Quantization=4-bit2025.06 | 0.297 | |
| RUTModel=Llama, Quantization=8-bit2025.06 | 0.253 | |
| MMDModel=Gemma, Quantization=4-bit2025.06 | 0.237 | |
| MMDModel=Mistral, Quantization=4-bit2025.06 | 0.216 | |
| MMDModel=Llama, Quantization=8-bit2025.06 | 0.177 | |
| KSModel=Mistral, Quantization=4-bit2025.06 | 0.16 | |
| KSModel=Llama, Quantization=8-bit2025.06 | 0.118 | |
| MMDModel=Gemma, Quantization=8-bit2025.06 | 0.067 | |
| MMDModel=Mistral, Quantization=8-bit2025.06 | 0.06 | |
| KSModel=Gemma, Quantization=4-bit2025.06 | 0.059 | |
| RUTModel=Mistral, Quantization=8-bit2025.06 | 0.059 | |
| RUTModel=Gemma, Quantization=8-bit2025.06 | 0.048 | |
| KSModel=Mistral, Quantization=8-bit2025.06 | 0.019 | |
| KSModel=Gemma, Quantization=8-bit2025.06 | 0.002 |