Wilcoxon T Tests on CLQA accuracy of GPLLMs with and without CLKR across C1-C8
| Knowledge domain | CLKR | Average accuracy | Accuracy enhancement | z-statistic | p-value |
|---|---|---|---|---|---|
| C1 | without | 0.466 | 14.5% | 6.672 | 0.000*** |
| with | 0.534 | ||||
| C2 | without | 0.449 | 28.2% | 5.896 | 0.000*** |
| with | 0.576 | ||||
| C3 | without | 0.449 | 21.6% | 6.825 | 0.000*** |
| with | 0.546 | ||||
| C4 | without | 0.408 | 21.1% | 7.086 | 0.000*** |
| with | 0.494 | ||||
| C5 | without | 0.458 | 24.5% | 5.966 | 0.000*** |
| with | 0.571 | ||||
| C6 | without | 0.465 | 23.6% | 7.427 | 0.000*** |
| with | 0.575 | ||||
| C7 | without | 0.462 | 21.6% | 7.334 | 0.000*** |
| with | 0.562 | ||||
| C8 | without | 0.470 | 17.9% | 5.970 | 0.000*** |
| with | 0.555 |
| Knowledge domain | CLKR | Average accuracy | Accuracy enhancement | ||
|---|---|---|---|---|---|
| C1 | without | 0.466 | 14.5% | 6.672 | 0.000*** |
| with | 0.534 | ||||
| C2 | without | 0.449 | 28.2% | 5.896 | 0.000*** |
| with | 0.576 | ||||
| C3 | without | 0.449 | 21.6% | 6.825 | 0.000*** |
| with | 0.546 | ||||
| C4 | without | 0.408 | 21.1% | 7.086 | 0.000*** |
| with | 0.494 | ||||
| C5 | without | 0.458 | 24.5% | 5.966 | 0.000*** |
| with | 0.571 | ||||
| C6 | without | 0.465 | 23.6% | 7.427 | 0.000*** |
| with | 0.575 | ||||
| C7 | without | 0.462 | 21.6% | 7.334 | 0.000*** |
| with | 0.562 | ||||
| C8 | without | 0.470 | 17.9% | 5.970 | 0.000*** |
| with | 0.555 |
Source(s): Authors’ own work