Multilingual Table Question Answering on MultiTEND ZH (Chinese)
23.47EMMultiLink
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| MultiLink2025.02 | 23.47 | 62.19 | 71.14 | 57.71 | 85.41 | 74.32 | |
| RAG for LLMEvaluation Protocol=RAG2025.02 | 13.98 | 59.52 | 67.86 | 47.02 | 73.6 | 68.47 | |
| SMART2025.02 | 13.94 | 57.69 | 60.36 | 38.05 | 83.1 | 75.85 | |
| Fine-tuned LlamaBackbone=Llama, Evaluation Protocol=Fine-tuned2025.02 | 13.57 | 56.71 | 53.83 | 36.86 | 78.84 | 66.28 | |
| Few-shot LLMEvaluation Protocol=Few-shot2025.02 | 10.25 | 58.52 | 63.32 | 34.95 | 63.79 | 63.21 | |
| Zero-shot LLMEvaluation Protocol=Zero-shot2025.02 | 0.61 | 47.76 | 58.59 | 28.99 | 54.4 | 57.47 |