Retrieval-Augmented Generation on NQ (Reliability Score (RS))
54.33Reliability Score (RS)CDA
Evaluation Results
| Method | Links | |
|---|---|---|
| CDABackbone=LLAMA3 8B INSTRUCT2024.12 | 54.33 | |
| CDA-MBackbone=LLAMA3 8B INSTRUCT2024.12 | 54.32 | |
| FSBBackbone=LLAMA3 8B INSTRUCT2024.12 | 53.66 | |
| CDA-MBackbone=LLAMA2 13B CHAT2024.12 | 51.95 | |
| ACD-ABackbone=LLAMA3 8B INSTRUCT2024.12 | 51.8 | |
| CDABackbone=LLAMA2 13B CHAT2024.12 | 51.42 | |
| ENTROPY (first-token)Backbone=LLAMA3 8B INSTRUCT2024.12 | 50.84 | |
| ABSTAINBackbone=LLAMA3 8B INSTRUCT2024.12 | 50.74 | |
| FSBBackbone=LLAMA2 13B CHAT2024.12 | 49.88 | |
| CDA-MBackbone=MISTRAL 7B INSTRUCT2024.12 | 49.51 | |
| ENTROPY (max)Backbone=LLAMA3 8B INSTRUCT2024.12 | 49.4 | |
| CDABackbone=MISTRAL 7B INSTRUCT2024.12 | 49.36 | |
| ABSTAINBackbone=MISTRAL 7B INSTRUCT2024.12 | 49.11 | |
| FSBBackbone=MISTRAL 7B INSTRUCT2024.12 | 49.05 | |
| ENTROPY (first-token)Backbone=MISTRAL 7B INSTRUCT2024.12 | 48.65 | |
| ENTROPY (max)Backbone=LLAMA2 13B CHAT2024.12 | 48.61 | |
| ENTROPY (average)Backbone=LLAMA3 8B INSTRUCT2024.12 | 48.46 | |
| ACD-ABackbone=MISTRAL 7B INSTRUCT2024.12 | 48.42 | |
| ENTROPY (max)Backbone=MISTRAL 7B INSTRUCT2024.12 | 48.38 | |
| ACD-ABackbone=LLAMA2 13B CHAT2024.12 | 48.3 | |
| ENTROPY (first-token)Backbone=LLAMA2 13B CHAT2024.12 | 48.28 | |
| ENTROPY (average)Backbone=LLAMA2 13B CHAT2024.12 | 47.63 | |
| CDABackbone=LLAMA2 7B CHAT2024.12 | 47.09 | |
| ENTROPY (average)Backbone=MISTRAL 7B INSTRUCT2024.12 | 46.84 | |
| CDA-MBackbone=LLAMA2 7B CHAT2024.12 | 46.4 | |
| SELF-ASKBackbone=LLAMA3 8B INSTRUCT2024.12 | 45.67 | |
| ENTROPY (min)Backbone=LLAMA3 8B INSTRUCT2024.12 | 44.76 | |
| ENTROPY (min)Backbone=LLAMA2 13B CHAT2024.12 | 43.73 | |
| ENTROPY (average)Backbone=LLAMA2 7B CHAT2024.12 | 42.88 | |
| ENTROPY (min)Backbone=MISTRAL 7B INSTRUCT2024.12 | 42.58 | |
| ENTROPY (first-token)Backbone=LLAMA2 7B CHAT2024.12 | 41.94 | |
| ABSTAINBackbone=LLAMA2 13B CHAT2024.12 | 41.54 | |
| ENTROPY (max)Backbone=LLAMA2 7B CHAT2024.12 | 41.36 | |
| ENTROPY (min)Backbone=LLAMA2 7B CHAT2024.12 | 41.3 | |
| SELF-ASKBackbone=MISTRAL 7B INSTRUCT2024.12 | 39.34 | |
| ACDBackbone=LLAMA3 8B INSTRUCT2024.12 | 38.4 | |
| FSBBackbone=LLAMA2 7B CHAT2024.12 | 37.36 | |
| ACDBackbone=LLAMA2 13B CHAT2024.12 | 35.65 | |
| SELF-ASKBackbone=LLAMA2 13B CHAT2024.12 | 34.15 | |
| ACD-ABackbone=LLAMA2 7B CHAT2024.12 | 33.56 | |
| CONTEXTBackbone=LLAMA3 8B INSTRUCT2024.12 | 33.35 | |
| ACDBackbone=LLAMA2 7B CHAT2024.12 | 33.02 | |
| ACDBackbone=MISTRAL 7B INSTRUCT2024.12 | 32.42 | |
| SELF-ASKBackbone=LLAMA2 7B CHAT2024.12 | 31.72 | |
| CONTEXTBackbone=LLAMA2 13B CHAT2024.12 | 31.15 | |
| CONTEXTBackbone=MISTRAL 7B INSTRUCT2024.12 | 30.68 | |
| CONTEXTBackbone=LLAMA2 7B CHAT2024.12 | 30.11 | |
| CADBackbone=LLAMA3 8B INSTRUCT2024.12 | 28.86 | |
| CADBackbone=MISTRAL 7B INSTRUCT2024.12 | 28.6 | |
| CADBackbone=LLAMA2 13B CHAT2024.12 | 28.4 | |
| CADBackbone=LLAMA2 7B CHAT2024.12 | 27.2 | |
| ABSTAINBackbone=LLAMA2 7B CHAT2024.12 | 22.06 |