Retrieval-Augmented Generation on ASQA
42.44str-EMConsJudge
Evaluation Results
| Method | Links | |
|---|---|---|
| ConsJudgeGenerator=Llama3-8B-Instruct, Reward Model Backbone=Llama3-8B-Instruct2025.02 | 42.44 | |
| Raw MetricGenerator=Llama3-8B-Instruct, Reward Model=Raw Metric2025.02 | 41.55 | |
| Vanilla LLMGenerator=Llama3-8B-Instruct, Reward Model Backbone=Llama3-8B-Instruct2025.02 | 40.69 | |
| SFTGenerator=Llama3-8B-Instruct, Reward Model Backbone=Llama3-8B-Instruct2025.02 | 40.37 | |
| ConsJudgeGenerator=MiniCPM-2.4B, Reward Model Backbone=Qwen2.5-14B-Instruct2025.02 | 36.45 | |
| Vanilla LLMGenerator=MiniCPM-2.4B, Reward Model Backbone=Llama3-8B-Instruct2025.02 | 35.77 | |
| ConsJudgeGenerator=MiniCPM-2.4B, Reward Model Backbone=Llama3-8B-Instruct2025.02 | 35.68 | |
| Vanilla LLMGenerator=MiniCPM-2.4B, Reward Model Backbone=Qwen2.5-14B-Instruct2025.02 | 34.95 | |
| SFTGenerator=MiniCPM-2.4B, Reward Model Backbone=Llama3-8B-Instruct2025.02 | 34.6 | |
| SFTGenerator=MiniCPM-2.4B, Reward Model Backbone=Qwen2.5-14B-Instruct2025.02 | 33.51 | |
| Raw MetricGenerator=MiniCPM-2.4B, Reward Model=Raw Metric2025.02 | 33.44 |