Tabular Question Answering on HiTab (val)
69.83Validation Reward (mean@4)ACPO
Evaluation Results
| Method | Links | |
|---|---|---|
| ACPOBackbone=Qwen2.5-7B, Policy Regime=N-OnP.2026.06 | 69.83 | |
| ACPOBackbone=Qwen2.5-7B, Policy Regime=N-OnP.2026.06 | 69.83 | |
| CISPOBackbone=Qwen2.5-7B, Policy Regime=N-OnP.2026.06 | 69.17 | |
| AR-LoptiBackbone=Qwen2.5-7B, Policy Regime=N-OnP.2026.06 | 68.92 | |
| DAPOBackbone=Qwen2.5-7B, Policy Regime=N-OnP.2026.06 | 68.42 | |
| Low-EntropyBackbone=Qwen2.5-7B, Policy Regime=N-OnP.2026.06 | 67 | |
| ACPOBackbone=Qwen2.5-7B, Policy Regime=OffP.2026.06 | 66.58 | |
| AR-LoptiBackbone=Qwen2.5-7B, Policy Regime=OffP.2026.06 | 66.58 | |
| ACPOBackbone=Qwen2.5-7B, Policy Regime=OffP.2026.06 | 66.58 | |
| DAPOBackbone=Qwen2.5-7B, Policy Regime=OffP.2026.06 | 65.75 | |
| CISPOBackbone=Qwen2.5-7B, Policy Regime=OffP.2026.06 | 65.67 | |
| Low-EntropyBackbone=Qwen2.5-7B, Policy Regime=OffP.2026.06 | 65.08 | |
| High-EntropyBackbone=Qwen2.5-7B, Policy Regime=N-OnP.2026.06 | 64.33 | |
| High-EntropyBackbone=Qwen2.5-7B, Policy Regime=OffP.2026.06 | 61.83 | |
| ACPOBackbone=Qwen2.5-3B, Policy Regime=OffP.2026.06 | 46.42 | |
| ACPOBackbone=Qwen2.5-3B, Policy Regime=OffP.2026.06 | 46.42 | |
| ACPOBackbone=Qwen2.5-3B, Policy Regime=N-OnP.2026.06 | 45.5 | |
| AR-LoptiBackbone=Qwen2.5-3B, Policy Regime=N-OnP.2026.06 | 45.5 | |
| ACPOBackbone=Qwen2.5-3B, Policy Regime=N-OnP.2026.06 | 45.5 | |
| High-EntropyBackbone=Qwen2.5-3B, Policy Regime=N-OnP.2026.06 | 45 | |
| CISPOBackbone=Qwen2.5-3B, Policy Regime=OffP.2026.06 | 44.83 | |
| DAPOBackbone=Qwen2.5-3B, Policy Regime=N-OnP.2026.06 | 44.33 | |
| DAPOBackbone=Qwen2.5-3B, Policy Regime=OffP.2026.06 | 44 | |
| CISPOBackbone=Qwen2.5-3B, Policy Regime=N-OnP.2026.06 | 43.92 | |
| High-EntropyBackbone=Qwen2.5-3B, Policy Regime=OffP.2026.06 | 43.5 | |
| AR-LoptiBackbone=Qwen2.5-3B, Policy Regime=OffP.2026.06 | 43.33 | |
| Low-EntropyBackbone=Qwen2.5-3B, Policy Regime=N-OnP.2026.06 | 43.17 | |
| Low-EntropyBackbone=Qwen2.5-3B, Policy Regime=OffP.2026.06 | 43 | |
| Base ModelBackbone=Qwen2.5-7B, Policy Regime=N-OnP.2026.06 | 31.25 | |
| Base ModelBackbone=Qwen2.5-7B2026.06 | 31.25 | |
| Base ModelBackbone=Qwen2.5-3B, Policy Regime=N-OnP.2026.06 | 17.5 | |
| Base ModelBackbone=Qwen2.5-3B2026.06 | 17.5 |