(self, messages, **kwargs)
| 19 | leaderboards: List[str] = field(default_factory=lambda: ["IFEval", "MATH", "GPQA", "MuSR", "MMLU", "Average"]) |
| 20 | |
| 21 | def generate(self, messages, **kwargs): |
| 22 | if self.model_name == 'DeepSeek-R1': |
| 23 | kwargs['temperature'] = 0.5 |
| 24 | return super().generate(messages, **kwargs) |
| 25 | |
| 26 | def step(self, tasks: List[Task], team: List[str],**kwargs): |
| 27 | """ |