Language Identification on WiLI 2018
100PrecisionSAEs Classifier
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| SAEs ClassifierBase Model=Llama 3.2 1B, Language=ko2025.07 | 100 | 99 | 99.5 | — | — | — | — | — | |
| SAEs ClassifierBase Model=Llama 3.2 1B, Language=vi2025.07 | 100 | 98.8 | 99.4 | — | — | — | — | — | |
| FastTextLanguage=bg2025.07 | 100 | 96.4 | 98.2 | — | — | — | — | — | |
| FastTextLanguage=hi2025.07 | 100 | 97.8 | 98.9 | — | — | — | — | — | |
| FastTextLanguage=ja2025.07 | 100 | 99.4 | 99.7 | — | — | — | — | — | |
| FastTextLanguage=ko2025.07 | 100 | 99 | 99.5 | — | — | — | — | — | |
| FastTextLanguage=th2025.07 | 100 | 98.8 | 99.4 | — | — | — | — | — | |
| FastTextLanguage=vi2025.07 | 100 | 98.6 | 99.3 | — | — | — | — | — | |
| Neurons ClassifierBase Model=Llama 3.2 1B, Language=th2025.07 | 100 | 97.2 | 98.6 | — | — | — | — | — | |
| Neurons ClassifierBase Model=Llama 3.2 1B, Language=vi2025.07 | 100 | 52 | 68.4 | — | — | — | — | — | |
| Neurons ClassifierBase Model=Llama 3.2 1B, Language=zh2025.07 | 100 | 0.2 | 0.4 | — | — | — | — | — | |
| SAEs ClassifierBase Model=Llama 3.2 1B, Language=ja2025.07 | 99.8 | 99.4 | 99.6 | — | — | — | — | — | |
| SAEs ClassifierBase Model=Llama 3.2 1B, Language=es2025.07 | 99.8 | 87.8 | 93.4 | — | — | — | — | — | |
| SAEs ClassifierBase Model=Llama 3.2 1B, Language=th2025.07 | 99.8 | 98.8 | 99.3 | — | — | — | — | — | |
| FastTextLanguage=pt2025.07 | 99.8 | 93.6 | 96.6 | — | — | — | — | — | |
| FastTextLanguage=it2025.07 | 99.6 | 96.6 | 98.1 | — | — | — | — | — | |
| FastTextLanguage=es2025.07 | 99.6 | 97.4 | 98.5 | — | — | — | — | — | |
| FastTextLanguage=tr2025.07 | 99.6 | 99.2 | 99.4 | — | — | — | — | — | |
| FastTextLanguage=ru2025.07 | 99.4 | 99.8 | 99.6 | — | — | — | — | — | |
| SAEs ClassifierBase Model=Llama 3.2 1B, Language=bg2025.07 | 99 | 96.6 | 97.8 | — | — | — | — | — | |
| SAEs ClassifierBase Model=Llama 3.2 1B, Language=hi2025.07 | 99 | 98.4 | 98.7 | — | — | — | — | — | |
| SAEs ClassifierBase Model=Llama 3.2 1B, Language=it2025.07 | 99 | 95.6 | 97.3 | — | — | — | — | — | |
| SAEs ClassifierBase Model=Llama 3.2 1B, Language=zh2025.07 | 99 | 99.6 | 99.3 | — | — | — | — | — | |
| FastTextLanguage=zh2025.07 | 99 | 99.6 | 99.3 | — | — | — | — | — | |
| SAEs ClassifierBase Model=Llama 3.2 1B, Language=ru2025.07 | 98.6 | 99.8 | 99.2 | — | — | — | — | — | |
| FastTextLanguage=fr2025.07 | 98.4 | 99.4 | 98.9 | — | — | — | — | — | |
| FastTextClassifier Type=FastText LID2025.07 | 98.4 | 98.3 | 98.3 | — | — | — | 98.3 | — | |
| fastTextClassifier Type=fastText2025.07 | 98.4 | 98.3 | 98.3 | — | — | — | 98.3 | — | |
| SAEs ClassifierBase Model=Llama 3.2 1B, Language=de2025.07 | 97.9 | 93.8 | 95.8 | — | — | — | — | — | |
| Neurons ClassifierBase Model=Llama 3.2 1B, Language=tr2025.07 | 97.6 | 99 | 98.3 | — | — | — | — | — | |
| SAEs ClassifierBase Model=Llama 3.2 1B, Language=fr2025.07 | 96.9 | 99.4 | 98.1 | — | — | — | — | — | |
| Neurons ClassifierBase Model=Llama 3.2 1B, Language=de2025.07 | 96.6 | 97.8 | 97.2 | — | — | — | — | — | |
| FastTextLanguage=de2025.07 | 96.5 | 98.4 | 97.4 | — | — | — | — | — | |
| Neurons ClassifierBase Model=Llama 3.2 1B, Language=fr2025.07 | 96.5 | 99 | 97.7 | — | — | — | — | — | |
| SAEs ClassifierBase Model=Llama 3.2 1B, Language=tr2025.07 | 95.9 | 99.4 | 97.6 | — | — | — | — | — | |
| Neurons ClassifierBase Model=Gemma 2 9B, Classifier Type=Neuron-based2025.07 | 94.6 | 92.2 | 90.5 | — | — | — | 92.2 | — | |
| SAEs ClassifierBase Model=Gemma 2 2B2025.07 | 94 | 92.4 | 90.6 | — | — | — | 92.4 | — | |
| SAEs ClassifierBase Model=Llama 3.2 1B, Language=pt2025.07 | 92.9 | 93.8 | 93.3 | — | — | — | — | — | |
| Neurons ClassifierBase Model=Gemma 2 2B2025.07 | 92.5 | 89.8 | 87.3 | — | — | — | 89.8 | — | |
| SAEs ClassifierBase Model=Gemma 2 9B, Classifier Type=SAE-based2025.07 | 88.9 | 91.7 | 89.7 | — | — | — | 91.7 | — | |
| Neurons ClassifierBase Model=Llama 3.2 1B, Language=it2025.07 | 86.6 | 95.6 | 90.9 | — | — | — | — | — | |
| FastTextLanguage=en2025.07 | 84.9 | 100 | 91.8 | — | — | — | — | — | |
| SAEs ClassifierBase Model=Llama 3.2 1B, Language=en2025.07 | 83.6 | 98 | 90.2 | — | — | — | — | — | |
| Neurons ClassifierBase Model=Llama 3.2 1B, Language=ja2025.07 | 54.3 | 97 | 69.6 | — | — | — | — | — | |
| Neurons ClassifierBase Model=Llama 3.2 1B, Language=hi2025.07 | 49.3 | 97.8 | 65.6 | — | — | — | — | — | |
| Neurons ClassifierBase Model=Llama 3.2 1B, Language=pt2025.07 | 49.3 | 93 | 64.4 | — | — | — | — | — | |
| Neurons ClassifierBase Model=Llama 3.2 1B, Language=bg2025.07 | 26.4 | 97.4 | 41.5 | — | — | — | — | — | |
| Neurons ClassifierBase Model=Llama 3.2 1B, Language=en2025.07 | 0 | 0 | 0 | — | — | — | — | — | |
| Neurons ClassifierBase Model=Llama 3.2 1B, Language=ko2025.07 | 0 | 0 | 0 | — | — | — | — | — | |
| Neurons ClassifierBase Model=Llama 3.2 1B, Language=ru2025.07 | 0 | 0 | 0 | — | — | — | — | — | |
| Neurons ClassifierBase Model=Llama 3.2 1B, Language=es2025.07 | 0 | 0 | 0 | — | — | — | — | — | |
| FastTextLanguage=Overall (Macro Average / Accuracy)2025.07 | — | — | — | 98.4 | 98.3 | 98.3 | 98.3 | — | |
| Neurons ClassifierBase Model=Llama 3.2 1B, Language=Overall (Macro Average / Accuracy)2025.07 | — | — | — | 57.1 | 61.7 | 52.8 | 61.7 | — | |
| SAEs ClassifierBase Model=Llama 3.2 1B, Language=Overall (Macro Average / Accuracy)2025.07 | — | — | — | 97.4 | 97.2 | 97.2 | 97.2 | — |