{"alpha":0.05,"base_model":"Qwen2.5-Coder-7B-Instruct Q4_K_M (QLoRA r16 a32 adapters)","baselines":{"astro_transfer":{"accuracy":0.7105,"correct":108,"n":152,"unparsable":0},"cyber":{"accuracy":0.65,"correct":325,"n":500,"unparsable":1},"exoplanet":{"accuracy":0.3617,"correct":217,"n":600,"unparsable":0},"maths":{"accuracy":0.785,"correct":157,"n":200,"unparsable":27},"physics_transfer":{"accuracy":0.541,"correct":264,"n":488,"unparsable":2},"stars":{"accuracy":0.489,"correct":155,"n":317,"unparsable":0}},"bonferroni_alpha":0.00294,"experiment":"NIGHT-MOE-1","n_other_cells":17,"subjects":{"cyber":{"control":{"accuracy":0.658,"base_accuracy_same_items":0.65,"correct":329,"delta_vs_base":0.008,"flag":"NS","mcnemar_exact_p":0.4545,"n":500,"only_adapter_right":10,"only_base_right":6,"unparsable":0,"wilson95":[0.6154,0.6982]},"control_note":"","missing_cells":["maths"],"other_packs":{"astro_transfer":{"accuracy":0.73,"base_accuracy_same_items":0.73,"correct":73,"delta_vs_base":0.0,"flag":"NS","kind":"interference","mcnemar_exact_p":1.0,"n":100,"only_adapter_right":1,"only_base_right":1,"survives_bonferroni":false,"unparsable":1,"wilson95":[0.6357,0.8073]},"exoplanet":{"accuracy":0.41,"base_accuracy_same_items":0.38,"correct":41,"delta_vs_base":0.03,"flag":"NS","kind":"interference","mcnemar_exact_p":0.25,"n":100,"only_adapter_right":3,"only_base_right":0,"survives_bonferroni":false,"unparsable":0,"wilson95":[0.3187,0.508]},"physics_transfer":{"accuracy":0.46,"base_accuracy_same_items":0.45,"correct":46,"delta_vs_base":0.01,"flag":"NS","kind":"interference","mcnemar_exact_p":1.0,"n":100,"only_adapter_right":3,"only_base_right":2,"survives_bonferroni":false,"unparsable":1,"wilson95":[0.3656,0.5574]},"stars":{"accuracy":0.54,"base_accuracy_same_items":0.46,"correct":54,"delta_vs_base":0.08,"flag":"GAIN","kind":"interference","mcnemar_exact_p":0.00781,"n":100,"only_adapter_right":8,"only_base_right":0,"survives_bonferroni":false,"unparsable":0,"wilson95":[0.4426,0.6344]}},"own":{"accuracy":0.66,"base_accuracy_same_items":0.65,"correct":330,"delta_vs_base":0.01,"flag":"NS","mcnemar_exact_p":0.38331,"n":500,"only_adapter_right":13,"only_base_right":8,"unparsable":1,"wilson95":[0.6174,0.7002]},"own_pack":"cyber","reason":"own-pack change +0.0100, p=0.38331","verdict":"NO_GAIN"},"exoplanet":{"control":{"accuracy":0.3867,"base_accuracy_same_items":0.3617,"correct":232,"delta_vs_base":0.025,"flag":"GAIN","mcnemar_exact_p":0.0237,"n":600,"only_adapter_right":27,"only_base_right":12,"unparsable":0,"wilson95":[0.3485,0.4262]},"control_note":"","missing_cells":["maths"],"other_packs":{"astro_transfer":{"accuracy":0.7171,"base_accuracy_same_items":0.7105,"correct":109,"delta_vs_base":0.0066,"flag":"NS","kind":"transfer","mcnemar_exact_p":1.0,"n":152,"only_adapter_right":1,"only_base_right":0,"survives_bonferroni":false,"unparsable":0,"wilson95":[0.6408,0.7827]},"cyber":{"accuracy":0.79,"base_accuracy_same_items":0.79,"correct":79,"delta_vs_base":0.0,"flag":"NS","kind":"interference","mcnemar_exact_p":1.0,"n":100,"only_adapter_right":0,"only_base_right":0,"survives_bonferroni":false,"unparsable":0,"wilson95":[0.7002,0.8583]},"physics_transfer":{"accuracy":0.41,"base_accuracy_same_items":0.45,"correct":41,"delta_vs_base":-0.04,"flag":"NS","kind":"interference","mcnemar_exact_p":0.125,"n":100,"only_adapter_right":0,"only_base_right":4,"survives_bonferroni":false,"unparsable":1,"wilson95":[0.3187,0.508]},"stars":{"accuracy":0.52,"base_accuracy_same_items":0.46,"correct":52,"delta_vs_base":0.06,"flag":"NS","kind":"interference","mcnemar_exact_p":0.07031,"n":100,"only_adapter_right":7,"only_base_right":1,"survives_bonferroni":false,"unparsable":0,"wilson95":[0.4232,0.6154]}},"own":{"accuracy":0.8733,"base_accuracy_same_items":0.3616,"correct":524,"delta_vs_base":0.5117,"flag":"GAIN","mcnemar_exact_p":0.0,"n":600,"only_adapter_right":324,"only_base_right":17,"unparsable":0,"wilson95":[0.8443,0.8976]},"own_pack":"exoplanet","reason":"shuffled-label control gained significantly over base (p=0.0237)","verdict":"VOID"},"maths":{"control":{"accuracy":0.0,"base_accuracy_same_items":0.75,"correct":0,"delta_vs_base":-0.75,"flag":"HARM","mcnemar_exact_p":0.0,"n":100,"only_adapter_right":0,"only_base_right":75,"unparsable":2,"wilson95":[0.0,0.037]},"control_note":"the shuffled control scored 0 correct, which is expected for exact-match numeric answers (a shuffled answer is almost never right), so this control cannot gain and the VOID rule is not testable for this subject","missing_cells":[],"other_packs":{"astro_transfer":{"accuracy":0.28,"base_accuracy_same_items":0.73,"correct":28,"delta_vs_base":-0.45,"flag":"HARM","kind":"interference","mcnemar_exact_p":0.0,"n":100,"only_adapter_right":1,"only_base_right":46,"survives_bonferroni":true,"unparsable":54,"wilson95":[0.2014,0.3749]},"cyber":{"accuracy":0.55,"base_accuracy_same_items":0.79,"correct":55,"delta_vs_base":-0.24,"flag":"HARM","kind":"interference","mcnemar_exact_p":0.0,"n":100,"only_adapter_right":0,"only_base_right":24,"survives_bonferroni":true,"unparsable":30,"wilson95":[0.4524,0.6439]},"exoplanet":{"accuracy":0.58,"base_accuracy_same_items":0.38,"correct":58,"delta_vs_base":0.2,"flag":"GAIN","kind":"interference","mcnemar_exact_p":0.0,"n":100,"only_adapter_right":20,"only_base_right":0,"survives_bonferroni":true,"unparsable":0,"wilson95":[0.4821,0.672]},"physics_transfer":{"accuracy":0.3422,"base_accuracy_same_items":0.541,"correct":167,"delta_vs_base":-0.1988,"flag":"HARM","kind":"transfer","mcnemar_exact_p":0.0,"n":488,"only_adapter_right":21,"only_base_right":118,"survives_bonferroni":true,"unparsable":155,"wilson95":[0.3015,0.3854]},"stars":{"accuracy":0.6,"base_accuracy_same_items":0.46,"correct":60,"delta_vs_base":0.14,"flag":"GAIN","kind":"interference","mcnemar_exact_p":0.00258,"n":100,"only_adapter_right":17,"only_base_right":3,"survives_bonferroni":true,"unparsable":0,"wilson95":[0.502,0.6906]}},"own":{"accuracy":0.715,"base_accuracy_same_items":0.785,"correct":143,"delta_vs_base":-0.07,"flag":"NS","mcnemar_exact_p":0.07039,"n":200,"only_adapter_right":19,"only_base_right":33,"unparsable":1,"wilson95":[0.6488,0.7731]},"own_pack":"maths","reason":"own-pack change -0.0700, p=0.07039","verdict":"NO_GAIN"},"stars":{"control":{"accuracy":0.306,"base_accuracy_same_items":0.489,"correct":97,"delta_vs_base":-0.183,"flag":"HARM","mcnemar_exact_p":5e-05,"n":317,"only_adapter_right":72,"only_base_right":130,"unparsable":0,"wilson95":[0.2578,0.3588]},"control_note":"","missing_cells":["maths"],"other_packs":{"astro_transfer":{"accuracy":0.7303,"base_accuracy_same_items":0.7106,"correct":111,"delta_vs_base":0.0197,"flag":"NS","kind":"transfer","mcnemar_exact_p":0.25,"n":152,"only_adapter_right":3,"only_base_right":0,"survives_bonferroni":false,"unparsable":0,"wilson95":[0.6547,0.7945]},"cyber":{"accuracy":0.79,"base_accuracy_same_items":0.79,"correct":79,"delta_vs_base":0.0,"flag":"NS","kind":"interference","mcnemar_exact_p":1.0,"n":100,"only_adapter_right":0,"only_base_right":0,"survives_bonferroni":false,"unparsable":0,"wilson95":[0.7002,0.8583]},"exoplanet":{"accuracy":0.38,"base_accuracy_same_items":0.38,"correct":38,"delta_vs_base":0.0,"flag":"NS","kind":"interference","mcnemar_exact_p":1.0,"n":100,"only_adapter_right":0,"only_base_right":0,"survives_bonferroni":false,"unparsable":0,"wilson95":[0.291,0.4779]},"physics_transfer":{"accuracy":0.44,"base_accuracy_same_items":0.45,"correct":44,"delta_vs_base":-0.01,"flag":"NS","kind":"interference","mcnemar_exact_p":1.0,"n":100,"only_adapter_right":2,"only_base_right":3,"survives_bonferroni":false,"unparsable":0,"wilson95":[0.3467,0.5377]}},"own":{"accuracy":0.8927,"base_accuracy_same_items":0.4889,"correct":283,"delta_vs_base":0.4038,"flag":"GAIN","mcnemar_exact_p":0.0,"n":317,"only_adapter_right":148,"only_base_right":20,"unparsable":0,"wilson95":[0.8539,0.9222]},"own_pack":"stars","reason":"own-pack gain +0.4038 (p=0.0); control HARM","verdict":"VALID"}}}