Next Occupation Prediction on Lightcast resume dataset 1,000 samples (test)
31.4Accuracy (EM)Qwen3-8B-Joint-SFT-DPO†
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen3-8B-Joint-SFT-DPO†Model Category=Proposed Models, Training Strategy=Joint-Model2026.04 | 31.4 | 36.46 | |
| Qwen3-8B-Two-Model-SFT-DPO†Model Category=Proposed Models, Training Strategy=Two-Model2026.04 | 31 | 35.35 | |
| Qwen3-8B-Joint-SFTModel Category=Proposed Models, Training Strategy=Joint-Model2026.04 | 28.2 | 32.83 | |
| LLaMA-3.1-8B-Joint-SFT-DPOModel Category=Proposed Models, Training Strategy=Joint-Model2026.04 | 27.6 | 32.85 | |
| Qwen3-8B-Two-Model-SFTModel Category=Proposed Models, Training Strategy=Two-Model2026.04 | 27.4 | 32.24 | |
| LLaMA-3.1-8B-Joint-SFTModel Category=Proposed Models, Training Strategy=Joint-Model2026.04 | 26.9 | 31.88 | |
| Qwen3-8B-Direct-SFTModel Category=No Reasoning2026.04 | 26.7 | 31.92 | |
| LLaMa-3.1-8B-Two-Model-SFT-DPOModel Category=Proposed Models, Training Strategy=Two-Model2026.04 | 26.7 | 31.62 | |
| LLaMa-3.1-8B-Two-Model-SFTModel Category=Proposed Models, Training Strategy=Two-Model2026.04 | 26.1 | 31.12 | |
| LLaMA-3.1-8B-Direct-SFTModel Category=No Reasoning2026.04 | 25.5 | 30.47 | |
| SASRecModel Category=Transformer Recommenders2026.04 | 25.3 | 29.83 | |
| Qwen3-8BModel Category=Zero-shot CoT2026.04 | 24.6 | 27.87 | |
| GPT-4.1Model Category=Zero-shot CoT2026.04 | 23.9 | 28.86 | |
| BERT4RecModel Category=Transformer Recommenders2026.04 | 23.6 | 28.82 | |
| LLaMA-3.1-8BModel Category=Zero-shot CoT2026.04 | 6.7 | 8.13 |