Multiple Choice Question Answering on TruthfulQA
34.64Accuracy (MC1 Delta)K-stage RK
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| K-stage RKSolver choice=Runge–Kutta, K=22026.05 | 34.64 | — | — | — | |
| Heun K=1Solver choice=Heun, K=12026.05 | 34.27 | — | — | — | |
| Baseline2026.05 | 34.15 | — | — | — | |
| DOLA dynamicModel=Qwen2.5-7B2026.06 | 0.2778 | — | — | 25.4 | |
| CAAModel=Mistral-7B, top_k=102026.06 | 0.2705 | — | — | 32.3 | |
| CAAModel=Qwen2.5-7B, top_k=102026.06 | 0.2558 | — | — | 15.5 | |
| CAAModel=Llama-3.1-8B, top_k=102026.06 | 0.2534 | — | — | 33.5 | |
| DOLA dynamicModel=Mistral-7B2026.06 | 0.2277 | — | — | 11.4 | |
| BaselineModel=Qwen2.5-7B2026.06 | 0.2215 | — | — | — | |
| BaselineModel=Mistral-7B2026.06 | 0.2044 | — | — | — | |
| DOLA dynamicModel=Llama-3.1-8B2026.06 | 0.1934 | — | — | 1.9 | |
| BaselineModel=Llama-3.1-8B2026.06 | 0.1897 | — | — | — | |
| DoLAModel=OPT-125M2026.03 | — | 0.03 | — | — | |
| DoLAModel=Phi-3-mini2026.03 | — | 0.07 | — | — | |
| DoLAModel=LLaMA 3-8B2026.03 | — | 0.08 | — | — | |
| H-Node ANCModel=LLaMA 3-8B2026.03 | — | 0.04 | — | — |