Intent Detection on Posts Dataset Out-of-Distribution
70.37F1 Score (val)self-consistency
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| self-consistencyUser Model=Llama-3.12026.05 | 70.37 | 75.34 | 60.03 | |
| BOT-MODUser Model=Llama-3.12026.05 | 70.13 | 79.43 | 52.47 | |
| zero-shot+User Model=Mistral-7B2026.05 | 69.98 | 71.41 | 66.78 | |
| zero-shotUser Model=Llama-3.12026.05 | 69.63 | 74.97 | 58.61 | |
| BOT-MODUser Model=Qwen32026.05 | 69.24 | 77.05 | 53.98 | |
| BERTUser Model=Qwen32026.05 | 67.91 | 71.49 | 60.23 | |
| zero-shot+User Model=Llama-3.12026.05 | 67.63 | 74.69 | 53.66 | |
| self-consistencyUser Model=Qwen32026.05 | 66.93 | 68.46 | 63.51 | |
| zero-shotUser Model=Qwen32026.05 | 65.94 | 67.42 | 62.63 | |
| self-consistencyUser Model=Mistral-7B2026.05 | 65.77 | 70.09 | 56.72 | |
| zero-shot+User Model=Qwen32026.05 | 65.66 | 68.63 | 59.23 | |
| zero-shotUser Model=Mistral-7B2026.05 | 65.2 | 69.46 | 56.29 | |
| self-refineUser Model=Llama-3.12026.05 | 64.61 | 71.59 | 51.74 | |
| self-refineUser Model=Qwen32026.05 | 64.51 | 66.15 | 61.06 | |
| BOT-MODUser Model=Mistral-7B2026.05 | 63.01 | 73.79 | 43.63 | |
| BERTUser Model=Llama-3.12026.05 | 62.47 | 66.35 | 54.29 | |
| CoTUser Model=Llama-3.12026.05 | 62.03 | 68.76 | 48.79 | |
| self-refineUser Model=Mistral-7B2026.05 | 61.91 | 63.95 | 57.46 | |
| CoTUser Model=Mistral-7B2026.05 | 61.27 | 65.07 | 53.24 | |
| CoTUser Model=Qwen32026.05 | 61.22 | 62.22 | 58.98 | |
| BERTUser Model=Mistral-7B2026.05 | 58.32 | 64.95 | 45.36 |