Temporal Reasoning (duration) on Synthetic Temporal Dataset (10-fold CV)
2.721Negative Log Sgemma-2-2b
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| gemma-2-2bManifold=log_lin2025.10 | 2.721 | 11 | |
| Qwen2.5-3B-ITManifold=log_lin2025.10 | 2.805 | 32 | |
| Llama-3.2-3BManifold=log_lin2025.10 | 2.879 | 21 | |
| Qwen2.5-3BManifold=log_lin2025.10 | 2.959 | 17 | |
| Llama-3.1-8B-ITManifold=log_lin2025.10 | 2.99 | 9 | |
| Llama-3.2-3B-ITManifold=log_lin2025.10 | 3.124 | 30 | |
| Llama-3.1-70B-ITManifold=log_lin2025.10 | 3.194 | 62 | |
| gemma-2-2b-ITManifold=log_lin2025.10 | 3.39 | 26 |