Temporal Reasoning (time_of_day_phase) on Synthetic Temporal Dataset 10-fold CV
6.182Log LossLlama-3.1-70B-IT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Llama-3.1-70B-ITManifold=clust2025.10 | 6.182 | 73 | |
| Llama-3.1-8B-ITManifold=clust2025.10 | 6.731 | 69 | |
| Llama-3.2-3BManifold=clust2025.10 | 7.061 | 64 | |
| Llama-3.2-3B-ITManifold=clust2025.10 | 7.072 | 66 | |
| gemma-2-2b-ITManifold=clust2025.10 | 7.589 | 29 | |
| gemma-2-2bManifold=clust2025.10 | 7.599 | 60 | |
| Qwen2.5-3BManifold=clust2025.10 | 8.711 | 45 | |
| Qwen2.5-3B-ITManifold=clust2025.10 | 8.803 | 30 |