Accuracy Evaluation on ROCStories (Writing)
76.3AccuracyES + TT-MV
Evaluation Results
| Method | Links | |
|---|---|---|
| ES + TT-MVModel=OLMo3-7B2026.03 | 76.3 | |
| ESModel=Llama3.1-8B-Inst2026.03 | 68.1 | |
| ES + TT-MVModel=OLMo3-7B-Inst2026.03 | 66 | |
| GRPOModel=OLMo3-7B-Inst2026.03 | 65.8 | |
| ESModel=OLMo3-7B-Inst2026.03 | 65.7 | |
| ES + TT-MVModel=Llama3.1-8B-Inst2026.03 | 65.2 | |
| PPOModel=OLMo3-7B2026.03 | 64.8 | |
| PPOModel=OLMo3-7B-Inst2026.03 | 64.7 | |
| ESModel=Qwen2.5-3B-Inst2026.03 | 64.5 | |
| RandOptModel=OLMo3-7B-Inst2026.03 | 64.5 | |
| RandOptModel=OLMo3-7B2026.03 | 64.5 | |
| ESModel=OLMo3-7B2026.03 | 64.4 | |
| BaseModel=OLMo3-7B-Inst2026.03 | 64 | |
| Best-of-N‡Model=OLMo3-7B-Inst2026.03 | 63.5 | |
| TT-MV†Model=OLMo3-7B-Inst2026.03 | 63.2 | |
| GRPOModel=OLMo3-7B2026.03 | 62.2 | |
| GRPOModel=Llama3.1-8B-Inst2026.03 | 62.2 | |
| ESModel=Qwen2.5-1.5B-Inst2026.03 | 60.2 | |
| ES + TT-MVModel=Qwen2.5-1.5B-Inst2026.03 | 59.1 | |
| RandOptModel=Llama3.1-8B-Inst2026.03 | 59 | |
| PPOModel=Llama3.1-8B-Inst2026.03 | 57.8 | |
| TT-MV†Model=Qwen2.5-3B-Inst2026.03 | 57.3 | |
| RandOptModel=Qwen2.5-3B-Inst2026.03 | 56.5 | |
| GRPOModel=Qwen2.5-3B-Inst2026.03 | 56.3 | |
| Best-of-N‡Model=Qwen2.5-3B-Inst2026.03 | 55 | |
| ES + TT-MVModel=Qwen2.5-3B-Inst2026.03 | 55 | |
| BaseModel=Qwen2.5-3B-Inst2026.03 | 54.7 | |
| Best-of-N‡Model=Llama3.1-8B-Inst2026.03 | 54 | |
| GRPOModel=Qwen2.5-1.5B-Inst2026.03 | 53.6 | |
| TT-MV†Model=Llama3.1-8B-Inst2026.03 | 53.2 | |
| PPOModel=Qwen2.5-1.5B-Inst2026.03 | 51.8 | |
| BaseModel=Llama3.1-8B-Inst2026.03 | 51.8 | |
| Best-of-N‡Model=OLMo3-7B2026.03 | 50 | |
| PPOModel=Qwen2.5-3B-Inst2026.03 | 49 | |
| RandOptModel=Qwen2.5-1.5B-Inst2026.03 | 48.5 | |
| BaseModel=Qwen2.5-1.5B-Inst2026.03 | 46.7 | |
| TT-MV†Model=Qwen2.5-1.5B-Inst2026.03 | 43.5 | |
| Best-of-N‡Model=Qwen2.5-1.5B-Inst2026.03 | 42.2 | |
| TT-MV†Model=OLMo3-7B2026.03 | 42.2 | |
| RandOptModel=Qwen2.5-0.5B-Inst2026.03 | 32.2 | |
| ESModel=Qwen2.5-0.5B-Inst2026.03 | 32 | |
| GRPOModel=Qwen2.5-0.5B-Inst2026.03 | 30.9 | |
| ES + TT-MVModel=Qwen2.5-0.5B-Inst2026.03 | 30.5 | |
| BaseModel=OLMo3-7B2026.03 | 24.1 | |
| BaseModel=Qwen2.5-0.5B-Inst2026.03 | 22 | |
| Best-of-N‡Model=Qwen2.5-0.5B-Inst2026.03 | 20.3 | |
| TT-MV†Model=Qwen2.5-0.5B-Inst2026.03 | 19.5 | |
| PPOModel=Qwen2.5-0.5B-Inst2026.03 | 19.1 |