Retrieval-Augmented Generation on TriviaQA (Reliability Score)
80.67Reliability ScoreCDA-M
Evaluation Results
| Method | Links | |
|---|---|---|
| CDA-MBackbone=LLAMA3 8B INSTRUCT2024.12 | 80.67 | |
| CDABackbone=LLAMA3 8B INSTRUCT2024.12 | 80.62 | |
| FSBBackbone=LLAMA3 8B INSTRUCT2024.12 | 79.14 | |
| CDA-MBackbone=LLAMA2 13B CHAT2024.12 | 78.74 | |
| CDABackbone=LLAMA2 13B CHAT2024.12 | 77.89 | |
| ABSTAINBackbone=LLAMA3 8B INSTRUCT2024.12 | 76.66 | |
| ENTROPY (first-token)Backbone=LLAMA3 8B INSTRUCT2024.12 | 76.2 | |
| ACD-ABackbone=LLAMA3 8B INSTRUCT2024.12 | 75.58 | |
| ENTROPY (max)Backbone=LLAMA2 13B CHAT2024.12 | 75.11 | |
| ENTROPY (first-token)Backbone=LLAMA2 13B CHAT2024.12 | 74.56 | |
| ENTROPY (max)Backbone=LLAMA3 8B INSTRUCT2024.12 | 74.24 | |
| ACDBackbone=LLAMA3 8B INSTRUCT2024.12 | 73.91 | |
| CDA-MBackbone=LLAMA2 7B CHAT2024.12 | 73.9 | |
| ENTROPY (average)Backbone=LLAMA2 13B CHAT2024.12 | 73.88 | |
| CDA-MBackbone=MISTRAL 7B INSTRUCT2024.12 | 73.22 | |
| CDABackbone=LLAMA2 7B CHAT2024.12 | 73.18 | |
| FSBBackbone=LLAMA2 7B CHAT2024.12 | 73.13 | |
| ABSTAINBackbone=MISTRAL 7B INSTRUCT2024.12 | 73 | |
| CDABackbone=MISTRAL 7B INSTRUCT2024.12 | 72.86 | |
| FSBBackbone=LLAMA2 13B CHAT2024.12 | 72.79 | |
| ENTROPY (average)Backbone=LLAMA3 8B INSTRUCT2024.12 | 72.12 | |
| ACDBackbone=LLAMA2 13B CHAT2024.12 | 72.09 | |
| ENTROPY (max)Backbone=LLAMA2 7B CHAT2024.12 | 71.88 | |
| ENTROPY (first-token)Backbone=LLAMA2 7B CHAT2024.12 | 71.87 | |
| ENTROPY (first-token)Backbone=MISTRAL 7B INSTRUCT2024.12 | 71.69 | |
| ENTROPY (max)Backbone=MISTRAL 7B INSTRUCT2024.12 | 71.57 | |
| ENTROPY (average)Backbone=MISTRAL 7B INSTRUCT2024.12 | 71.54 | |
| ABSTAINBackbone=LLAMA2 7B CHAT2024.12 | 70.7 | |
| ACD-ABackbone=LLAMA2 7B CHAT2024.12 | 70.55 | |
| ACD-ABackbone=MISTRAL 7B INSTRUCT2024.12 | 70.39 | |
| FSBBackbone=MISTRAL 7B INSTRUCT2024.12 | 70.26 | |
| ENTROPY (min)Backbone=LLAMA2 13B CHAT2024.12 | 70.18 | |
| ENTROPY (average)Backbone=LLAMA2 7B CHAT2024.12 | 70.12 | |
| ENTROPY (min)Backbone=LLAMA3 8B INSTRUCT2024.12 | 68.61 | |
| SELF-ASKBackbone=MISTRAL 7B INSTRUCT2024.12 | 67.94 | |
| SELF-ASKBackbone=LLAMA2 13B CHAT2024.12 | 67.73 | |
| ACDBackbone=LLAMA2 7B CHAT2024.12 | 67.35 | |
| CONTEXTBackbone=LLAMA2 13B CHAT2024.12 | 66.81 | |
| ACDBackbone=MISTRAL 7B INSTRUCT2024.12 | 66.24 | |
| ENTROPY (min)Backbone=MISTRAL 7B INSTRUCT2024.12 | 66.19 | |
| ACD-ABackbone=LLAMA2 13B CHAT2024.12 | 65.78 | |
| ENTROPY (min)Backbone=LLAMA2 7B CHAT2024.12 | 65.67 | |
| CONTEXTBackbone=LLAMA3 8B INSTRUCT2024.12 | 65.3 | |
| SELF-ASKBackbone=LLAMA3 8B INSTRUCT2024.12 | 64.91 | |
| SELF-ASKBackbone=LLAMA2 7B CHAT2024.12 | 63.12 | |
| CONTEXTBackbone=MISTRAL 7B INSTRUCT2024.12 | 62.4 | |
| CADBackbone=LLAMA2 13B CHAT2024.12 | 61.8 | |
| CONTEXTBackbone=LLAMA2 7B CHAT2024.12 | 61.01 | |
| CADBackbone=LLAMA3 8B INSTRUCT2024.12 | 59.32 | |
| CADBackbone=MISTRAL 7B INSTRUCT2024.12 | 56.09 | |
| CADBackbone=LLAMA2 7B CHAT2024.12 | 54.44 | |
| ABSTAINBackbone=LLAMA2 13B CHAT2024.12 | 47.11 |