CN
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Multi-Agent Reinforcement Learning | CN rac-dist | Mean Episodic Reward888 | 21 | |
| Multi-Agent Reinforcement Learning | CN rdist | Mean Episodic Reward-161 | 21 | |
| Multi-Agent Reinforcement Learning | CN rdete | Mean Episodic Reward-154 | 21 | |
| Role-Playing Evaluation (Conversational-Naturalness) | CN | Win Rate65 | 9 | |
| CN task stream | CN (Medium) | Backward Transfer35 | 8 | |
| CN task stream | CN (Expert) | Backward Transfer5.83 | 8 | |
| Multi-agent Continual Cooperation | CN Medium | Forward Transfer16 | 7 | |
| Multi-agent Continual Cooperation | CN Expert | Forward Transfer6.74 | 7 | |
| Cooperative Navigation | CN MPE hard | Mean Episode Reward3.37 | 7 | |
| Cooperative Navigation | CN MPE medium | Mean Episode Reward3.21 | 7 | |
| Soft Query Answering | CN15k | 1P Score16.6 | 6 | |
| Backdoor Attack | CN (test) | Runtime (s)28.3 | 4 | |
| Intent Prediction | CN | Accuracy55.2 | 4 | |
| Function Invocation | CN Ver. Dual | Token Usage1,377.9 | 3 | |
| Function Invocation | CN (Single) | Invocation Accuracy0.89 | 3 | |
| Competitive ratio of ε-LDP online (Wε) vs. ε-LDP offline (Lε) stopping | Cn independent non-negative random variables | Lower Bound Competitive Ratio1 | 1 | |
| Competitive ratio of ε-LDP online (Wε) vs. non-private offline (M) stopping | Cn independent non-negative random variables | Lower Bound Competitive Ratio2 | 1 | |
| Competitive ratio of ε-LDP online (Wε) vs. non-private online (V) stopping | Cn independent non-negative random variables | Lower Bound Competitive Ratio1 | 1 | |
| Competitive ratio of non-private online (V) vs. non-private offline (M) stopping | Cn independent non-negative random variables | Lower Bound Competitive Ratio1 | 1 |
Showing 19 of 19 rows