Probabilistic Risk Prediction on SEC risk (test)
0.1979Brier ScoreQwen3-32B
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Qwen3-32BTraining Phase=Foresight Learning fine-tuned, Checkpoint=step 1442026.01 | 0.1979 | 11.6 | 0.0287 | |
| GPT-52026.01 | 0.1986 | 11.3 | 0.0812 | |
| Naive Baseline2026.01 | 0.2238 | 0 | — | |
| Qwen3-32BTraining Phase=pretrained base model2026.01 | 0.2381 | -6.4 | 0.1419 |