Timeline Summarization on Crisis
10.8Align R-1TimelineReasoner
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| TimelineReasonerModel=QwQ-32B2026.04 | 10.8 | 3.31 | 35.09 | |
| NTS-CoT2026.06 | 9.4 | 2.4 | 30.5 | |
| One-shotPrompting=One-shot2026.06 | 9.1 | 2.2 | 27.1 | |
| DATEWISE2026.06 | 8.9 | 2.6 | 29.5 | |
| LLM-TLS2026.06 | 8.9 | 2.2 | 28.8 | |
| CHRONOSModel=Qwen2.5-32B2026.04 | 8.61 | 2.95 | 25.83 | |
| SDF2026.06 | 8.6 | 1.8 | 30.2 | |
| CHRONOSModel=QwQ-32B2026.04 | 8.28 | 2.56 | 29.07 | |
| STEEN2026.06 | 8.2 | 2.7 | 29.7 | |
| EGC2026.04 | 7.91 | 1.49 | 29.11 | |
| EGC2026.06 | 7.9 | 1.5 | 29.1 | |
| MARTSCHAT2026.06 | 7.5 | 1.6 | 28.1 | |
| LLM-TLSModel=Qwen2.5-32B2026.04 | 7.01 | 1.99 | 27.32 | |
| Zero-shotPrompting=Zero-shot2026.06 | 6.8 | 1.6 | 27.1 | |
| LLM-TLSModel=QwQ-32B2026.04 | 6.78 | 1.89 | 27 | |
| CLUST2026.04 | 6.1 | 1.31 | 22.6 | |
| CLUST2026.06 | 6.1 | 1.3 | 22.6 | |
| Standard-CotPrompting=Chain-of-Thought2026.06 | 5.9 | 1.4 | 25.7 | |
| ITER_RAGModel=QwQ-32B2026.04 | 5.79 | 1.73 | 22.01 | |
| ITER_RAGModel=Qwen2.5-32B2026.04 | 5.52 | 1.3 | 12.79 | |
| REWRITEModel=QwQ-32B2026.04 | 4.41 | 2.43 | 10.3 | |
| REWRITEModel=Qwen2.5-32B2026.04 | 3.42 | 1.9 | 9.27 | |
| DIRECTModel=QwQ-32B2026.04 | 1.84 | 0.2 | 6.78 | |
| DIRECTModel=Qwen2.5-32B2026.04 | 1.62 | 0.4 | 6.39 |