SPB Git

spb/wp26_uqo

Public
2.0 KB · 61 lines json
1{2 "generated_at": "2026-10-28T08:45:51+00:00",3 "script": "scripts/20_evaluate.py",4 "runs": [5  "baselines",6  "final_E_single_llm_mixed_random_mid",7  "final_F_independent_mixed_random_mid",8  "final_G_debate_r2_mixed_random_mid",9  "final_H_debate_skeptic_r2_mixed_random_mid",10  "final_I_placebo_r2_mixed_random_mid",11  "final_J_cost_matched_mixed_random_mid",12  "final_small_E_single_llm_mixed_random_small",13  "final_small_F_independent_mixed_random_small",14  "final_small_G_debate_r2_mixed_random_small",15  "holdout_E_single_llm_mixed_random_mid",16  "holdout_F_independent_mixed_random_mid",17  "holdout_G_debate_r2_mixed_random_mid",18  "prospective_E_single_llm_mixed_random_mid",19  "prospective_F_independent_mixed_random_mid",20  "prospective_G_debate_r2_mixed_random_mid",21  "prospective_baselines",22  "repeat_G_debate_r2_mixed_random_mid",23  "validation_E_single_llm_all_anthropic_mid",24  "validation_E_single_llm_all_openai_mid",25  "validation_E_single_llm_mixed_random_mid",26  "validation_F_independent_all_anthropic_mid",27  "validation_F_independent_all_openai_mid",28  "validation_F_independent_mixed_random_mid",29  "validation_G_debate_r1_mixed_random_mid",30  "validation_G_debate_r2_all_anthropic_mid",31  "validation_G_debate_r2_all_openai_mid",32  "validation_G_debate_r2_mixed_random_mid",33  "validation_G_debate_r4_mixed_random_mid",34  "validation_G_fullview_r2_mixed_random_mid",35  "validation_H_debate_skeptic_r2_mixed_random_mid",36  "validation_I_placebo_r2_mixed_random_mid",37  "validation_J_cost_matched_mixed_random_mid"38 ],39 "truth": "/Users/simon-pierreboucher/Desktop/Recherche-UQO/immobilier/agentic-valuation-quebec/data/processed/analytic.parquet",40 "split": "holdout",41 "n_prediction_rows": 502194,42 "n_truth": 83489,43 "B": 2000,44 "seed": 20261027,45 "main_committee": "mixed_random",46 "main_aggregation": "median",47 "min_cell": 30,48 "arms": [49  "A_hedonic",50  "B_hierarchical",51  "C_catboost",52  "C_gbm",53  "C_xgboost",54  "D_avm",55  "E_single_llm",56  "F_independent",57  "G_debate_r2"58 ],59 "fake_included": false,60 "note": ""61}
Files · main
Menu

Theme