SPB Git

Repository Public Research preview

summary.md 3.8 KB

system n chrf edit_sim rougeL exact_match number_recall placeholders_exact
B0_identity 3000 68.423 [67.595, 69.232] 0.678 [0.670, 0.686] 0.684 [0.675, 0.693] 0.000 [0.000, 0.000] 1.000 [1.000, 1.000] 1.000 [1.000, 1.000]
B1_generic_paraphrase 3000 65.671 [64.912, 66.414] 0.566 [0.558, 0.574] 0.600 [0.591, 0.610] 0.000 [0.000, 0.000] 1.000 [0.999, 1.000] 1.000 [1.000, 1.000]
B2_t5-small 3000 68.382 [67.598, 69.139] 0.680 [0.672, 0.688] 0.687 [0.679, 0.696] 0.000 [0.000, 0.000] 0.989 [0.983, 0.994] 0.713 [0.666, 0.759]
htr-edit-tiny 3000 66.154 [65.271, 67.034] 0.671 [0.663, 0.679] 0.702 [0.693, 0.711] 0.000 [0.000, 0.000] 1.000 [1.000, 1.000] 1.000 [1.000, 1.000]
htr-m1-small-mediumdenoiseinit 3000 68.619 [67.819, 69.398] 0.675 [0.668, 0.683] 0.684 [0.675, 0.692] 0.000 [0.000, 0.000] 0.990 [0.987, 0.993] 0.996 [0.993, 0.998]
htr-m1-small-v2 3000 68.290 [67.517, 69.047] 0.673 [0.665, 0.681] 0.681 [0.673, 0.689] 0.000 [0.000, 0.000] 0.987 [0.983, 0.990] 0.993 [0.989, 0.996]
htr-m1-tiny-mediumdenoiseinit 3000 68.749 [67.942, 69.538] 0.678 [0.670, 0.686] 0.686 [0.678, 0.695] 0.000 [0.000, 0.000] 0.991 [0.988, 0.994] 0.992 [0.987, 0.996]
htr-m1-tiny-pilotdenoiseinit 3000 67.873 [67.105, 68.624] 0.674 [0.666, 0.681] 0.675 [0.667, 0.683] 0.000 [0.000, 0.000] 0.985 [0.980, 0.989] 0.988 [0.983, 0.994]
htr-m3tokenonly-tiny-mediumdenoiseinit-s2 3000 68.735 [67.937, 69.501] 0.679 [0.671, 0.686] 0.688 [0.680, 0.697] 0.000 [0.000, 0.000] 0.990 [0.986, 0.994] 0.991 [0.986, 0.996]
htr-m3tokenonly-tiny-mediumdenoiseinit 3000 68.776 [67.963, 69.573] 0.680 [0.673, 0.688] 0.689 [0.680, 0.697] 0.000 [0.000, 0.000] 0.990 [0.986, 0.993] 0.995 [0.991, 0.998]
htr-medium_m1_tiny_mediumdenoiseinit_s2 3000 68.807 [68.000, 69.602] 0.679 [0.672, 0.687] 0.688 [0.679, 0.697] 0.000 [0.000, 0.000] 0.992 [0.989, 0.994] 0.994 [0.990, 0.998]
htr-medium_m1_tiny_mediumdenoiseinit_s3 3000 68.841 [68.033, 69.615] 0.679 [0.671, 0.686] 0.687 [0.679, 0.696] 0.000 [0.000, 0.000] 0.991 [0.988, 0.993] 0.995 [0.992, 0.998]
htr-medium_m2_tiny_mediumdenoiseinit 3000 68.572 [67.770, 69.351] 0.677 [0.669, 0.684] 0.685 [0.677, 0.694] 0.000 [0.000, 0.000] 0.988 [0.985, 0.992] 0.995 [0.991, 0.999]
htr-medium_m3light_tiny_mediumdenoiseinit 3000 68.484 [67.664, 69.271] 0.676 [0.668, 0.683] 0.684 [0.676, 0.693] 0.000 [0.000, 0.000] 0.991 [0.987, 0.993] 0.994 [0.990, 0.998]

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

  • rewrite vs original: 0.164
  • B0_identity vs original: 0.164 · probe P(original) mean 0.42
  • B1_generic_paraphrase vs original: 0.226 · probe P(original) mean 0.45
  • B2_t5-small vs original: 0.117 · probe P(original) mean 0.49
  • htr-edit-tiny vs original: 0.157 · probe P(original) mean 0.51
  • htr-m1-small-mediumdenoiseinit vs original: 0.095 · probe P(original) mean 0.53
  • htr-m1-small-v2 vs original: 0.099 · probe P(original) mean 0.52
  • htr-m1-tiny-mediumdenoiseinit vs original: 0.100 · probe P(original) mean 0.52
  • htr-m1-tiny-pilotdenoiseinit vs original: 0.126 · probe P(original) mean 0.47
  • htr-m3tokenonly-tiny-mediumdenoiseinit-s2 vs original: 0.105 · probe P(original) mean 0.51
  • htr-m3tokenonly-tiny-mediumdenoiseinit vs original: 0.106 · probe P(original) mean 0.51
  • htr-medium_m1_tiny_mediumdenoiseinit_s2 vs original: 0.097 · probe P(original) mean 0.52
  • htr-medium_m1_tiny_mediumdenoiseinit_s3 vs original: 0.100 · probe P(original) mean 0.52
  • htr-medium_m2_tiny_mediumdenoiseinit vs original: 0.098 · probe P(original) mean 0.52
  • htr-medium_m3light_tiny_mediumdenoiseinit vs original: 0.098 · probe P(original) mean 0.53
Files main
Menu

Theme