Automated Probing on MMLU
46Error Rate (%)PAIR
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PAIRGenerator Model=GPT-5.2, Target Model=Llama-3.1-8b-ins.2026.02 | 46 | 94.87 | |
| AutoDetectGenerator Model=GPT-5.2, Target Model=Llama-3.1-8b-ins.2026.02 | 67 | — | |
| PROBELLMGenerator Model=GPT-5.2, Target Model=Llama-3.1-8b-ins.2026.02 | 86 | — |