Common Sense Reasoning on HELLASWAG 1800
53.8AccuracyQwen3-4B
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3-4B2026.06 | 53.8 | |
| Talkie-1930variant=base2026.06 | 52.2 | |
| GPT-2 XL2026.06 | 41.1 | |
| TYPEWRITERLMvariant=base2026.06 | 39 | |
| GPT-1900variant=base2026.06 | 36 |
| Method | Links | |
|---|---|---|
| Qwen3-4B2026.06 | 53.8 | |
| Talkie-1930variant=base2026.06 | 52.2 | |
| GPT-2 XL2026.06 | 41.1 | |
| TYPEWRITERLMvariant=base2026.06 | 39 | |
| GPT-1900variant=base2026.06 | 36 |