Natural Language Inference on ANLI R1
78.5AccuracySelf-consistency
Evaluation results
| Method | Links | |
|---|---|---|
| Self-consistencyPrompting strategy=Self-consistency, Backbone model=PaLM-540B2022.03 | 78.5 | |
| PaLM 2-Lprompting=1-shot2023.05 | 73.1 | |
| Standard-prompting (no-rationale)Prompting strategy=no-rationale, Backbone model=PaLM-540B2022.03 | 69.1 | |
| CoT-promptingPrompting strategy=Chain-of-Thought, Backbone model=PaLM-540B2022.03 | 68.8 | |
| PaLM 2-Mprompting=1-shot2023.05 | 58.1 | |
| PaLM 2-Sprompting=1-shot2023.05 | 53.1 | |
| PaLMprompting=1-shot2023.05 | 52.6 |