### qwen3-8b · medial · geo
   idiom config:    {'model': 'Qwen/Qwen3-8B', 'reduction': 'geometric_mean', 'medial_only': True, 'dtype': 'bfloat16', 'dataset': '/home/prada/PID_evaluation/data/dataset.tsv', 'num_idioms': 18, 'syn_reg_eps': 0.01}
   nonidiom config: {'model': 'Qwen/Qwen3-8B', 'reduction': 'geometric_mean', 'medial_only': True, 'dtype': 'bfloat16', 'dataset': '/home/prada/PID_evaluation/data/nonidioms_dataset.tsv', 'num_idioms': 18, 'syn_reg_eps': 0.01}

== idioms :: ratio_u_idiom ==  (N=18 phrases)
     mean     median                    95% CI
   1.2015     1.1701  [  1.1675,  1.2372]

== non-idioms :: ratio_u_idiom ==  (N=18 phrases)
     mean     median                    95% CI
   1.0719     1.0652  [  1.0536,  1.0903]
  cross-dataset ratio_u_idiom: idioms - nonidioms  Δ=+0.1296  CI=[+0.0907,+0.1700]  *

== idioms :: ratio_s_idiom ==  (N=15 phrases)  (3 non-finite dropped)
     mean     median                    95% CI
   1.1528     1.1306  [  1.1282,  1.1802]

== non-idioms :: ratio_s_idiom ==  (N=6 phrases)  (12 non-finite dropped)
     mean     median                    95% CI
   1.2560     1.2585  [  1.2121,  1.2940]
  cross-dataset ratio_s_idiom: idioms - nonidioms  Δ=-0.1032  CI=[-0.1510,-0.0536]  *