gsaltintas commited on
Commit
43772d5
·
verified ·
1 Parent(s): 38fb5fd

Uploading tokenizer_robustness_completion_italian_english_keyboard subset

Browse files
README.md CHANGED
@@ -36,6 +36,10 @@ configs:
36
  data_files:
37
  - split: test
38
  path: tokenizer_robustness_completion_italian_dialects/test-*
 
 
 
 
39
  dataset_info:
40
  - config_name: tokenizer_robustness_completion_italian_abbreviations
41
  features:
@@ -919,6 +923,132 @@ dataset_info:
919
  num_examples: 31
920
  download_size: 35296
921
  dataset_size: 16845
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
922
  ---
923
 
924
  # Dataset Card for Tokenization Robustness
 
36
  data_files:
37
  - split: test
38
  path: tokenizer_robustness_completion_italian_dialects/test-*
39
+ - config_name: tokenizer_robustness_completion_italian_english_keyboard
40
+ data_files:
41
+ - split: test
42
+ path: tokenizer_robustness_completion_italian_english_keyboard/test-*
43
  dataset_info:
44
  - config_name: tokenizer_robustness_completion_italian_abbreviations
45
  features:
 
923
  num_examples: 31
924
  download_size: 35296
925
  dataset_size: 16845
926
+ - config_name: tokenizer_robustness_completion_italian_english_keyboard
927
+ features:
928
+ - name: question
929
+ dtype: string
930
+ - name: choices
931
+ list: string
932
+ - name: answer
933
+ dtype: int64
934
+ - name: answer_label
935
+ dtype: string
936
+ - name: split
937
+ dtype: string
938
+ - name: subcategories
939
+ dtype: string
940
+ - name: category
941
+ dtype: string
942
+ - name: lang
943
+ dtype: string
944
+ - name: second_lang
945
+ dtype: string
946
+ - name: notes
947
+ dtype: string
948
+ - name: id
949
+ dtype: string
950
+ - name: set_id
951
+ dtype: string
952
+ - name: variation_id
953
+ dtype: string
954
+ - name: perturbed_word
955
+ dtype: string
956
+ - name: vanilla_cos_sim_to_canonical
957
+ struct:
958
+ - name: CohereLabs/aya-expanse-8b
959
+ dtype: float64
960
+ - name: Qwen/Qwen3-8B
961
+ dtype: float64
962
+ - name: bigscience/bloom
963
+ dtype: float64
964
+ - name: common-pile/comma-v0.1-1t
965
+ dtype: float64
966
+ - name: facebook/xglm-564M
967
+ dtype: float64
968
+ - name: google-bert/bert-base-multilingual-cased
969
+ dtype: float64
970
+ - name: google/byt5-small
971
+ dtype: float64
972
+ - name: google/gemma-2-2b
973
+ dtype: float64
974
+ - name: gpt2
975
+ dtype: float64
976
+ - name: meta-llama/Llama-3.2-1B
977
+ dtype: float64
978
+ - name: microsoft/Phi-3-mini-4k-instruct
979
+ dtype: float64
980
+ - name: mistralai/tekken
981
+ dtype: float64
982
+ - name: tiktoken/gpt-4o
983
+ dtype: float64
984
+ - name: tokenmonster/englishcode-32000-consistent-v1
985
+ dtype: float64
986
+ - name: trimmed_cos_sim_to_canonical
987
+ struct:
988
+ - name: CohereLabs/aya-expanse-8b
989
+ dtype: float64
990
+ - name: Qwen/Qwen3-8B
991
+ dtype: float64
992
+ - name: bigscience/bloom
993
+ dtype: float64
994
+ - name: common-pile/comma-v0.1-1t
995
+ dtype: float64
996
+ - name: facebook/xglm-564M
997
+ dtype: float64
998
+ - name: google-bert/bert-base-multilingual-cased
999
+ dtype: float64
1000
+ - name: google/byt5-small
1001
+ dtype: float64
1002
+ - name: google/gemma-2-2b
1003
+ dtype: float64
1004
+ - name: gpt2
1005
+ dtype: float64
1006
+ - name: meta-llama/Llama-3.2-1B
1007
+ dtype: float64
1008
+ - name: microsoft/Phi-3-mini-4k-instruct
1009
+ dtype: float64
1010
+ - name: mistralai/tekken
1011
+ dtype: float64
1012
+ - name: tiktoken/gpt-4o
1013
+ dtype: float64
1014
+ - name: tokenmonster/englishcode-32000-consistent-v1
1015
+ dtype: float64
1016
+ - name: token_counts
1017
+ struct:
1018
+ - name: CohereLabs/aya-expanse-8b
1019
+ dtype: int64
1020
+ - name: Qwen/Qwen3-8B
1021
+ dtype: int64
1022
+ - name: bigscience/bloom
1023
+ dtype: int64
1024
+ - name: common-pile/comma-v0.1-1t
1025
+ dtype: int64
1026
+ - name: facebook/xglm-564M
1027
+ dtype: int64
1028
+ - name: google-bert/bert-base-multilingual-cased
1029
+ dtype: int64
1030
+ - name: google/byt5-small
1031
+ dtype: int64
1032
+ - name: google/gemma-2-2b
1033
+ dtype: int64
1034
+ - name: gpt2
1035
+ dtype: int64
1036
+ - name: meta-llama/Llama-3.2-1B
1037
+ dtype: int64
1038
+ - name: microsoft/Phi-3-mini-4k-instruct
1039
+ dtype: int64
1040
+ - name: mistralai/tekken
1041
+ dtype: int64
1042
+ - name: tiktoken/gpt-4o
1043
+ dtype: int64
1044
+ - name: tokenmonster/englishcode-32000-consistent-v1
1045
+ dtype: int64
1046
+ splits:
1047
+ - name: test
1048
+ num_bytes: 36901
1049
+ num_examples: 68
1050
+ download_size: 41603
1051
+ dataset_size: 36901
1052
  ---
1053
 
1054
  # Dataset Card for Tokenization Robustness
tokenizer_robustness_completion_italian_english_keyboard/test-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cafc061990a6c53fe1630aac33f24bdf1cb1e26a9a544509027804d3f63efeae
3
+ size 41603