Table 4. The accuracy of ChatGPT on the GAT Arabic and English analogy questions compared to using fastText embeddings.
The best performance is indicated in bold.
| Language | ChatGPT | fastText |
|---|---|---|
| Arabic | 37.19% | 43.80% |
| English | 54.03% | 33.87% |
The best performance is indicated in bold.
| Language | ChatGPT | fastText |
|---|---|---|
| Arabic | 37.19% | 43.80% |
| English | 54.03% | 33.87% |