SPB Git

Repository Public Research preview

summary.md 1.6 KB

system n chrf edit_sim rougeL exact_match number_recall placeholders_exact
B0_identity 812 69.895 [68.945, 71.069] 0.660 [0.647, 0.672] 0.670 [0.658, 0.684] 0.000 [0.000, 0.000] 1.000 [1.000, 1.000] 1.000 [1.000, 1.000]
B1_generic_paraphrase 812 67.223 [66.195, 68.451] 0.562 [0.551, 0.573] 0.596 [0.585, 0.610] 0.000 [0.000, 0.000] 0.999 [0.997, 1.000] 1.000 [1.000, 1.000]
htr-m1-small 812 70.273 [69.403, 71.293] 0.663 [0.651, 0.674] 0.675 [0.664, 0.688] 0.000 [0.000, 0.000] 0.990 [0.982, 0.996] 0.994 [0.986, 0.999]
htr-m1-tiny-s3 812 70.618 [69.785, 71.580] 0.669 [0.657, 0.681] 0.680 [0.669, 0.692] 0.000 [0.000, 0.000] 0.989 [0.982, 0.995] 0.995 [0.985, 1.000]
htr-m1-tiny 812 70.661 [69.843, 71.621] 0.669 [0.658, 0.680] 0.681 [0.670, 0.692] 0.000 [0.000, 0.000] 0.991 [0.986, 0.995] 0.989 [0.975, 0.997]
htr-m3tokenonly-tiny 812 70.517 [69.663, 71.516] 0.668 [0.657, 0.679] 0.680 [0.669, 0.692] 0.000 [0.000, 0.000] 0.993 [0.988, 0.997] 0.990 [0.970, 1.000]

Stylometric W1 (mean over features, standardised; lower = closer to originals):

  • rewrite vs original: 0.180
  • B0_identity vs original: 0.180 · probe P(original) mean 0.39
  • B1_generic_paraphrase vs original: 0.328 · probe P(original) mean 0.36
  • htr-m1-small vs original: 0.114 · probe P(original) mean 0.51
  • htr-m1-tiny-s3 vs original: 0.119 · probe P(original) mean 0.50
  • htr-m1-tiny vs original: 0.116 · probe P(original) mean 0.50
  • htr-m3tokenonly-tiny vs original: 0.130 · probe P(original) mean 0.48
Files main
Menu

Theme