False Refusal Evaluation on False Refusal Benchmarks
77OKTestSelf-CD
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Self-CDdecoding_settings=greedy decoding, temperature=none, max_length=5122026.03 | 77 | 42.6 | 78 | |
| VCLdecoding_settings=greedy decoding, temperature=none, max_length=5122026.03 | 76 | 87 | 86 | |
| System Promptdecoding_settings=greedy decoding, temperature=none, max_length=5122026.03 | 71 | 65 | 58 | |
| Vector Ablationdecoding_settings=greedy decoding, temperature=none, max_length=5122026.03 | 67 | 44.7 | 58 | |
| DROdecoding_settings=greedy decoding, temperature=none, max_length=5122026.03 | 63 | 71 | 68 | |
| Llama-3.2-1B-SFTdecoding_settings=greedy decoding, temperature=none, max_length=5122026.03 | 53 | 76 | 51 |