Question Answering on LogiQA
44.29AccuracyDS2-INSTRUCT
Evaluation results
| Method | Links | |
|---|---|---|
| DS2-INSTRUCTModel Family=Qwen2.52026.03 | 44.29 | |
| InstructMixModel Family=Qwen2.52026.03 | 43.88 | |
| Self-InstructModel Family=Qwen2.52026.03 | 42.56 | |
| ExploreInstructModel Family=Qwen2.52026.03 | 41.19 | |
| DS2-INSTRUCTModel Family=Llama32026.03 | 35.33 | |
| Zero-ShotModel Family=Qwen2.52026.03 | 32.77 | |
| DS2-INSTRUCTModel Family=Mistral2026.03 | 32.23 | |
| Self-InstructModel Family=Mistral2026.03 | 29.67 | |
| InstructMixModel Family=Mistral2026.03 | 26.72 | |
| ExploreInstructModel Family=Mistral2026.03 | 25.18 | |
| Self-InstructModel Family=Llama32026.03 | 24.16 | |
| GPT-JParameters=6B, Evaluation=Five-shot2022.04 | 23 | |
| ExploreInstructModel Family=Llama32026.03 | 22.35 | |
| GPT-NeoXParameters=20B, Evaluation=Five-shot2022.04 | 22 | |
| InstructMixModel Family=Llama32026.03 | 15.88 | |
| Zero-ShotModel Family=Llama32026.03 | 9.77 | |
| Zero-ShotModel Family=Mistral2026.03 | 9.52 |