Commonsense Question Answering on ECQA (test)
79.7AccuracyLlama2-70B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Llama2-70BPrompt Order=predict-then-explain (PE)2024.04 | 79.7 | 24.1 | 0.083 | |
| Llama2-70BPrompt Order=explain-then-predict (EP)2024.04 | 77.8 | 28.8 | 0.038 | |
| NV-EmbedBase Model=Llama-3.1-8B, Evaluation Protocol=Fine-tuned, Number of Seeds=52026.05 | 71.75 | — | — | |
| Llama2-13BPrompt Order=explain-then-predict (EP)2024.04 | 71.4 | 30.2 | 0.036 | |
| LESSBase Model=Llama-3.1-8B, Evaluation Protocol=Fine-tuned, Number of Seeds=52026.05 | 71.19 | — | — | |
| Interaction-Aware Influence Function (Ours)Base Model=Llama-3.1-8B, Evaluation Protocol=Fine-tuned, Number of Seeds=52026.05 | 71.17 | — | — | |
| RDS+Base Model=Llama-3.1-8B, Evaluation Protocol=Fine-tuned, Number of Seeds=52026.05 | 70.88 | — | — | |
| RandomBase Model=Llama-3.1-8B, Evaluation Protocol=Fine-tuned, Number of Seeds=52026.05 | 69.85 | — | — | |
| Additive IFBase Model=Llama-3.1-8B, Evaluation Protocol=Fine-tuned, Number of Seeds=52026.05 | 69.52 | — | — | |
| Llama2-13BPrompt Order=predict-then-explain (PE)2024.04 | 68 | 28.6 | 0.055 | |
| Llama2-7BPrompt Order=explain-then-predict (EP)2024.04 | 55.2 | 31.7 | 0.065 | |
| Llama2-7BPrompt Order=predict-then-explain (PE)2024.04 | 54.1 | 30.4 | 0.047 | |
| Random2024.04 | 20 | — | 0 |