Temporal Reasoning on Date Understanding
89.53AccuracyGRPO
Evaluation results
| Method | Links | |
|---|---|---|
| GRPOBase Model=Qwen3-4B-Instruct-25072026.04 | 89.53 | |
| Cog-DRIFTBase Model=Qwen3-4B-Instruct-25072026.04 | 88.75 | |
| Few-shotBase Model=Qwen3-4B-Instruct-25072026.04 | 88.37 | |
| Zero-shotBase Model=Qwen3-4B-Instruct-25072026.04 | 87.71 | |
| NuRL (Abstract)Base Model=Qwen3-4B-Instruct-25072026.04 | 87.48 | |
| RFTBase Model=Qwen3-4B-Instruct-25072026.04 | 85.11 | |
| NuRL (Prefix)Base Model=Qwen3-4B-Instruct-25072026.04 | 84.17 | |
| Cog-DRIFTBase Model=Llama3.2-3B-Instruct2026.04 | 62.82 | |
| NuRL (Abstract)Base Model=Llama3.2-3B-Instruct2026.04 | 56.72 | |
| Few-shotBase Model=Llama3.2-3B-Instruct2026.04 | 55.63 | |
| NuRL (Prefix)Base Model=Llama3.2-3B-Instruct2026.04 | 54.91 | |
| Zero-shotBase Model=Llama3.2-3B-Instruct2026.04 | 54.87 | |
| RFTBase Model=Llama3.2-3B-Instruct2026.04 | 50.31 | |
| GRPOBase Model=Llama3.2-3B-Instruct2026.04 | 38.85 |