Anomaly Detection on TSRBench
91.67AccuracyChatTS
Evaluation Results
| Method | Links | |
|---|---|---|
| ChatTSBase Model=Qwen2.5-3B-Instruct, Training Protocol=SFT2026.02 | 91.67 | |
| VeriTimeBase Model=Qwen3-4B-Instruct, Training Protocol=SFT+RL2026.02 | 91.11 | |
| VeriTimeBase Model=Qwen2.5-3B-Instruct, Training Protocol=SFT+RL2026.02 | 90.56 | |
| ChatTSBase Model=Qwen3-4B-Instruct, Training Protocol=SFT2026.02 | 89.44 | |
| GPT-4o-miniModel Type=General LLM2026.02 | 82.12 | |
| Qwen3-4B-InstructTraining Protocol=Base2026.02 | 77.78 | |
| Mistral-7B-v0.3Model Type=General LLM2026.02 | 75 | |
| Qwen2.5-7B-instructModel Type=General LLM2026.02 | 70.56 | |
| Time-MQABase Model=Mistral-7B2026.02 | 69.81 | |
| Meta-Llama3-8B-InstructModel Type=General LLM2026.02 | 68.89 | |
| Time-MQABase Model=Llama3-8B2026.02 | 64.15 | |
| Time-R1Base Model=Qwen2.5-7B2026.02 | 54.55 | |
| DeepSeek-R1-Distill-Qwen-7BModel Type=General LLM2026.02 | 48.33 | |
| Time-MQABase Model=Qwen2.5-7B2026.02 | 32.22 | |
| Qwen2.5-3B-InstructTraining Protocol=Base2026.02 | 26.67 |