Narrative Writing on ROCStories (test)
65.8AccuracyGRPO
Evaluation results
| Method | Links | |
|---|---|---|
| GRPOModel=OLMo3-7B-Inst2026.05 | 65.8 | |
| CoRPModel=OLMo3-7B-Inst2026.05 | 64.8 | |
| PPOModel=OLMo3-7B-Inst2026.05 | 64.7 | |
| RandOptModel=OLMo3-7B-Inst, K=502026.05 | 64.5 | |
| RandOptModel=OLMo3-7B-Inst, K=12026.05 | 64.2 | |
| BaseModel=OLMo3-7B-Inst2026.05 | 64 | |
| GRPOModel=Llama3.1-8B-Inst2026.05 | 62.2 | |
| CoRPModel=Llama3.1-8B-Inst2026.05 | 60.5 | |
| CoRPModel=Qwen2.5-3B-Inst2026.05 | 59.1 | |
| RandOptModel=Llama3.1-8B-Inst, K=502026.05 | 59 | |
| PPOModel=Llama3.1-8B-Inst2026.05 | 57.8 | |
| RandOptModel=Qwen2.5-3B-Inst, K=502026.05 | 56.5 | |
| GRPOModel=Qwen2.5-3B-Inst2026.05 | 56.3 | |
| BaseModel=Qwen2.5-3B-Inst2026.05 | 54.7 | |
| RandOptModel=Qwen2.5-3B-Inst, K=12026.05 | 54.6 | |
| GRPOModel=Qwen2.5-1.5B-Inst2026.05 | 53.6 | |
| CoRPModel=Qwen2.5-1.5B-Inst2026.05 | 52.2 | |
| PPOModel=Qwen2.5-1.5B-Inst2026.05 | 51.8 | |
| BaseModel=Llama3.1-8B-Inst2026.05 | 51.8 | |
| RandOptModel=Llama3.1-8B-Inst, K=12026.05 | 50.8 | |
| PPOModel=Qwen2.5-3B-Inst2026.05 | 49 | |
| RandOptModel=Qwen2.5-1.5B-Inst, K=502026.05 | 48.5 | |
| BaseModel=Qwen2.5-1.5B-Inst2026.05 | 46.7 | |
| RandOptModel=Qwen2.5-1.5B-Inst, K=12026.05 | 46.6 | |
| RandOptModel=Qwen2.5-0.5B-Inst, K=502026.05 | 32.2 | |
| CoRPModel=Qwen2.5-0.5B-Inst2026.05 | 32 | |
| GRPOModel=Qwen2.5-0.5B-Inst2026.05 | 30.9 | |
| RandOptModel=Qwen2.5-0.5B-Inst, K=12026.05 | 23.3 | |
| BaseModel=Qwen2.5-0.5B-Inst2026.05 | 22 | |
| PPOModel=Qwen2.5-0.5B-Inst2026.05 | 19.1 |