RAG Performance Prediction on TriviaQA
0.889QE5 ScoreBert-Gen
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Bert-GenLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | 0.889 | 0.738 | 0.704 | |
| Bert-GenLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | 0.885 | 0.664 | 0.642 | |
| Bert-GenLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | 0.873 | 0.694 | 0.653 | |
| Bert-GenLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | 0.868 | 0.655 | 0.554 | |
| UncertaintyLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | 0.564 | 0.474 | 0.427 | |
| UncertaintyLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | 0.547 | 0.283 | 0.296 | |
| UncertaintyLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | 0.545 | 0.428 | 0.388 | |
| UncertaintyLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | 0.512 | 0.236 | 0.254 | |
| Bert-PostLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | 0.46 | 0.456 | 0.445 | |
| Bert-PostLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | 0.304 | 0.425 | 0.403 | |
| Bert-PostLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | 0.296 | 0.282 | 0.298 | |
| REFLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | 0.286 | 0.226 | 0.236 | |
| EntailmentLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | 0.229 | 0.159 | 0.174 | |
| REFLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | 0.214 | 0.205 | 0.19 | |
| REFLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | 0.161 | 0.127 | 0.126 | |
| maxVARLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | 0.16 | 0.021 | 0.035 | |
| NQCLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | 0.159 | 0.142 | 0.117 | |
| QCLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | 0.155 | 0.14 | 0.114 | |
| QCLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | 0.151 | 0.12 | 0.103 | |
| U_SMVLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | 0.15 | 0.118 | 0.099 | |
| WIGLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | 0.144 | 0.107 | 0.115 | |
| NQCLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | 0.141 | 0.161 | 0.175 | |
| SMVLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | 0.138 | 0.157 | 0.173 | |
| QCLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | 0.134 | 0.154 | 0.167 | |
| QCLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | 0.132 | 0.115 | 0.122 | |
| U_SMVLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | 0.13 | 0.114 | 0.122 | |
| U_WIGLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | 0.127 | 0.101 | 0.105 | |
| Bert-PostLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | 0.121 | 0.319 | 0.33 | |
| U_SMVLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | 0.117 | 0.116 | 0.13 | |
| U_WIGLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | 0.109 | 0.067 | 0.044 | |
| REFLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | 0.105 | 0.094 | 0.115 | |
| U_WIGLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | 0.103 | 0.071 | 0.07 | |
| EntailmentLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | 0.096 | 0.026 | 0.05 | |
| maxIDFLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | 0.076 | 0.089 | 0.087 | |
| NQCLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | 0.074 | 0.066 | 0.087 | |
| maxVARLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | 0.065 | 0.076 | 0.077 | |
| EntailmentLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | 0.064 | 0.061 | 0.079 | |
| WIGLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | 0.061 | 0.036 | 0.063 | |
| maxSCQLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | 0.057 | 0.07 | 0.066 | |
| SMVLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | 0.053 | 0.042 | 0.065 | |
| SMVLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | 0.05 | 0.058 | 0.053 | |
| U_SMVLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | 0.049 | 0.057 | 0.051 | |
| meanVARLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | 0.046 | 0.061 | 0.069 | |
| meanIDFLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | 0.042 | 0.063 | 0.073 | |
| maxSCQLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | 0.039 | 0.04 | 0.047 | |
| meanSCQLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | 0.033 | 0.052 | 0.066 | |
| NQCLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | 0.023 | 0.013 | 0.024 | |
| SMVLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | 0.022 | 0.013 | 0.024 | |
| WIGLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | 0.019 | -0.007 | -0.019 | |
| U_WIGLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | 0.016 | -0.025 | 0.004 | |
| maxIDFLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | 0.015 | 0.018 | 0.026 | |
| maxSCQLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | 0.01 | 0.023 | 0.027 | |
| maxVARLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | 0.006 | 0.014 | 0.026 | |
| maxSCQLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | 0.002 | 0.017 | 0.018 | |
| maxIDFLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | -0.004 | 0.005 | 0.023 | |
| meanSCQLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | -0.004 | 0.005 | 0.014 | |
| meanVARLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | -0.011 | -0.011 | 0.001 | |
| maxVARLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | -0.013 | -0.004 | 0.006 | |
| EntailmentLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | -0.013 | -0.004 | 0.035 | |
| meanSCQLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | -0.019 | -0.009 | 0.022 | |
| meanIDFLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | -0.021 | -0.013 | 0.005 | |
| meanVARLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | -0.023 | -0.011 | 0.013 | |
| maxIDFLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | -0.025 | -0.018 | -0.012 | |
| minSCQLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | -0.029 | -0.024 | -0.011 | |
| minIDFLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | -0.03 | -0.025 | -0.01 | |
| minVARLLM=Llama, Retrieval Method=BM25 Retrieval2026.04 | -0.033 | -0.026 | -0.008 | |
| minVARLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | -0.039 | -0.022 | -0.006 | |
| minSCQLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | -0.044 | -0.043 | -0.025 | |
| meanIDFLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | -0.045 | -0.029 | 0 | |
| minVARLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | -0.048 | -0.056 | -0.048 | |
| minVARLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | -0.048 | -0.049 | -0.043 | |
| minIDFLLM=Llama, Retrieval Method=E5 Retrieval2026.04 | -0.049 | -0.045 | -0.026 | |
| meanSCQLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | -0.053 | -0.034 | -0.016 | |
| WIGLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | -0.054 | -0.059 | -0.015 | |
| meanVARLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | -0.058 | -0.041 | -0.022 | |
| minSCQLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | -0.06 | -0.064 | -0.043 | |
| minIDFLLM=Falcon, Retrieval Method=BM25 Retrieval2026.04 | -0.067 | -0.07 | -0.049 | |
| minSCQLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | -0.069 | -0.075 | -0.05 | |
| minIDFLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | -0.078 | -0.082 | -0.056 | |
| meanIDFLLM=Falcon, Retrieval Method=E5 Retrieval2026.04 | -0.081 | -0.063 | -0.043 |