Tabular Reasoning on MultiHierTT
28.7Pass@1 AccuracyTAC
Evaluation Results
| Method | Links | |
|---|---|---|
| TACBackbone=Qwen3-1.7B2026.06 | 28.7 | |
| SECBackbone=Qwen3-1.7B2026.06 | 26.5 | |
| RandomBackbone=Qwen3-1.7B2026.06 | 25.5 | |
| M2OBackbone=Qwen3-1.7B2026.06 | 24.8 | |
| TACBackbone=Llama3.2-3B2026.06 | 24.8 | |
| M2OBackbone=Llama3.2-3B2026.06 | 23 | |
| RandomBackbone=Llama3.2-3B2026.06 | 21.3 | |
| SECBackbone=Llama3.2-3B2026.06 | 17.8 | |
| BaseBackbone=Llama3.2-3B2026.06 | 16.6 | |
| BaseBackbone=Qwen3-1.7B2026.06 | 14.7 |