Science Reasoning on ARC Challenge (test)
84.04AccuracyGenius
Evaluation results
| Method | Links | |
|---|---|---|
| GeniusTraining Corpus=Magpie (25K), Supervision Level=Without Supervision, Backbone Model=LLaMA3.1-8B-Instruct2025.04 | 84.04 | |
| SPINTraining Corpus=Magpie (25K), Supervision Level=With Supervision, Backbone Model=LLaMA3.1-8B-Instruct2025.04 | 83.96 | |
| GeniusTraining Corpus=OpenHermes2.5 (32K), Supervision Level=Without Supervision, Backbone Model=LLaMA3.1-8B-Instruct2025.04 | 83.19 | |
| CoHTraining Corpus=Magpie (25K), Supervision Level=Without Supervision, Backbone Model=LLaMA3.1-8B-Instruct2025.04 | 82.51 | |
| Self-RewardingTraining Corpus=Magpie (25K), Supervision Level=Without Supervision, Backbone Model=LLaMA3.1-8B-Instruct2025.04 | 82.17 | |
| Self-RewardingTraining Corpus=OpenHermes2.5 (32K), Supervision Level=Without Supervision, Backbone Model=LLaMA3.1-8B-Instruct2025.04 | 81.66 | |
| CoHTraining Corpus=OpenHermes2.5 (32K), Supervision Level=Without Supervision, Backbone Model=LLaMA3.1-8B-Instruct2025.04 | 81.48 | |
| SCPOTraining Corpus=OpenHermes2.5 (32K), Supervision Level=Without Supervision, Backbone Model=LLaMA3.1-8B-Instruct2025.04 | 79.52 | |
| SCPOTraining Corpus=Magpie (25K), Supervision Level=Without Supervision, Backbone Model=LLaMA3.1-8B-Instruct2025.04 | 78.92 | |
| LLaMA3.1-8B-InstructReasoning Strategy=CoT, Backbone Model=LLaMA3.1-8B-Instruct2025.04 | 78.33 | |
| SFTTraining Corpus=Magpie (25K), Supervision Level=With Supervision, Backbone Model=LLaMA3.1-8B-Instruct2025.04 | 74.06 | |
| SPINTraining Corpus=OpenHermes2.5 (32K), Supervision Level=With Supervision, Backbone Model=LLaMA3.1-8B-Instruct2025.04 | 71.76 | |
| SFTTraining Corpus=OpenHermes2.5 (32K), Supervision Level=With Supervision, Backbone Model=LLaMA3.1-8B-Instruct2025.04 | 69.37 | |
| STaRTraining Corpus=OpenHermes2.5 (32K), Supervision Level=Without Supervision, Backbone Model=LLaMA3.1-8B-Instruct2025.04 | 68.43 | |
| STaRTraining Corpus=Magpie (25K), Supervision Level=Without Supervision, Backbone Model=LLaMA3.1-8B-Instruct2025.04 | 67.24 |