Robust Safety and Utility Evaluation in Federated Learning on MaliciousGen & LMSYS-Chat
92.5Rule Compliance
Shadow-Level
Updated 3mo ago
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
2026.01 | 92.5 | 79.04 | -1.49 | 3.34 | |
2026.01 | 91.15 | 78.08 | -1.64 | 3.14 | |
2026.01 | 90.96 | 74.42 | -1.54 | 4.01 | |
2026.01 | 61.73 | 15.38 | -3.56 | 3.22 | |
2026.01 | 52.88 | 12.31 | -3.79 | 3.14 | |
2026.01 | 52.88 | 10.77 | -3.84 | 3.06 | |
2026.01 | 52.12 | 10.58 | -3.81 | 2.98 | |
2026.01 | 51.15 | 12.31 | -3.64 | 3.03 |