Classification on Titanic (val)
0.883AUROCHuman-LLM Collaborative Feature Engineering Framework
Evaluation results
| Method | Links | |
|---|---|---|
| Human-LLM Collaborative Feature Engineering FrameworkDownstream Model=XGBoost, LLM Backbone=GPT-4o, Human Involvement=Active2026.01 | 0.883 | |
| Human-LLM Collaborative Feature Engineering FrameworkDownstream Model=XGBoost, LLM Backbone=GPT-4o, Human Involvement=Absent2026.01 | 0.879 | |
| OCTreeDownstream Model=XGBoost, LLM Backbone=GPT-4o2026.01 | 0.874 | |
| Human-LLM Collaborative Feature Engineering FrameworkDownstream Model=MLP, LLM Backbone=GPT-4o, Human Involvement=Active2026.01 | 0.87 | |
| CAAFEDownstream Model=XGBoost, LLM Backbone=GPT-4o2026.01 | 0.87 | |
| Human-LLM Collaborative Feature Engineering FrameworkDownstream Model=MLP, LLM Backbone=GPT-4o, Human Involvement=Absent2026.01 | 0.868 | |
| OCTreeDownstream Model=MLP, LLM Backbone=GPT-4o2026.01 | 0.865 | |
| CAAFEDownstream Model=MLP, LLM Backbone=GPT-4o2026.01 | 0.863 | |
| OpenFEDownstream Model=XGBoost2026.01 | 0.85 | |
| AutoGluonDownstream Model=XGBoost2026.01 | 0.848 | |
| AutoGluonDownstream Model=MLP2026.01 | 0.843 | |
| OpenFEDownstream Model=MLP2026.01 | 0.841 |