gsaltintas commited on
Commit
76f61ef
·
verified ·
1 Parent(s): 502ccd2

Uploading tokenizer_robustness_completion_english_keyboard_proximity_errors subset

Browse files
README.md CHANGED
@@ -1868,6 +1868,130 @@ dataset_info:
1868
  num_examples: 14
1869
  download_size: 31848
1870
  dataset_size: 7330
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1871
  configs:
1872
  - config_name: tokenizer_robustness_completion_english_abbreviations
1873
  data_files:
@@ -1929,6 +2053,10 @@ configs:
1929
  data_files:
1930
  - split: test
1931
  path: tokenizer_robustness_completion_english_inflections/test-*
 
 
 
 
1932
  ---
1933
 
1934
  # Dataset Card for Tokenization Robustness
 
1868
  num_examples: 14
1869
  download_size: 31848
1870
  dataset_size: 7330
1871
+ - config_name: tokenizer_robustness_completion_english_keyboard_proximity_errors
1872
+ features:
1873
+ - name: question
1874
+ dtype: string
1875
+ - name: choices
1876
+ list: string
1877
+ - name: answer
1878
+ dtype: int64
1879
+ - name: answer_label
1880
+ dtype: string
1881
+ - name: split
1882
+ dtype: string
1883
+ - name: subcategories
1884
+ dtype: string
1885
+ - name: category
1886
+ dtype: string
1887
+ - name: lang
1888
+ dtype: string
1889
+ - name: second_lang
1890
+ dtype: string
1891
+ - name: notes
1892
+ dtype: string
1893
+ - name: id
1894
+ dtype: string
1895
+ - name: set_id
1896
+ dtype: string
1897
+ - name: variation_id
1898
+ dtype: string
1899
+ - name: vanilla_cos_sim_to_canonical
1900
+ struct:
1901
+ - name: CohereLabs/aya-expanse-8b
1902
+ dtype: float64
1903
+ - name: Qwen/Qwen3-8B
1904
+ dtype: float64
1905
+ - name: bigscience/bloom
1906
+ dtype: float64
1907
+ - name: common-pile/comma-v0.1-1t
1908
+ dtype: float64
1909
+ - name: facebook/xglm-564M
1910
+ dtype: float64
1911
+ - name: google-bert/bert-base-multilingual-cased
1912
+ dtype: float64
1913
+ - name: google/byt5-small
1914
+ dtype: float64
1915
+ - name: google/gemma-2-2b
1916
+ dtype: float64
1917
+ - name: gpt2
1918
+ dtype: float64
1919
+ - name: meta-llama/Llama-3.2-1B
1920
+ dtype: float64
1921
+ - name: microsoft/Phi-3-mini-4k-instruct
1922
+ dtype: float64
1923
+ - name: mistralai/tekken
1924
+ dtype: float64
1925
+ - name: tiktoken/gpt-4o
1926
+ dtype: float64
1927
+ - name: tokenmonster/englishcode-32000-consistent-v1
1928
+ dtype: float64
1929
+ - name: trimmed_cos_sim_to_canonical
1930
+ struct:
1931
+ - name: CohereLabs/aya-expanse-8b
1932
+ dtype: float64
1933
+ - name: Qwen/Qwen3-8B
1934
+ dtype: float64
1935
+ - name: bigscience/bloom
1936
+ dtype: float64
1937
+ - name: common-pile/comma-v0.1-1t
1938
+ dtype: float64
1939
+ - name: facebook/xglm-564M
1940
+ dtype: float64
1941
+ - name: google-bert/bert-base-multilingual-cased
1942
+ dtype: float64
1943
+ - name: google/byt5-small
1944
+ dtype: float64
1945
+ - name: google/gemma-2-2b
1946
+ dtype: float64
1947
+ - name: gpt2
1948
+ dtype: float64
1949
+ - name: meta-llama/Llama-3.2-1B
1950
+ dtype: float64
1951
+ - name: microsoft/Phi-3-mini-4k-instruct
1952
+ dtype: float64
1953
+ - name: mistralai/tekken
1954
+ dtype: float64
1955
+ - name: tiktoken/gpt-4o
1956
+ dtype: float64
1957
+ - name: tokenmonster/englishcode-32000-consistent-v1
1958
+ dtype: float64
1959
+ - name: token_counts
1960
+ struct:
1961
+ - name: CohereLabs/aya-expanse-8b
1962
+ dtype: int64
1963
+ - name: Qwen/Qwen3-8B
1964
+ dtype: int64
1965
+ - name: bigscience/bloom
1966
+ dtype: int64
1967
+ - name: common-pile/comma-v0.1-1t
1968
+ dtype: int64
1969
+ - name: facebook/xglm-564M
1970
+ dtype: int64
1971
+ - name: google-bert/bert-base-multilingual-cased
1972
+ dtype: int64
1973
+ - name: google/byt5-small
1974
+ dtype: int64
1975
+ - name: google/gemma-2-2b
1976
+ dtype: int64
1977
+ - name: gpt2
1978
+ dtype: int64
1979
+ - name: meta-llama/Llama-3.2-1B
1980
+ dtype: int64
1981
+ - name: microsoft/Phi-3-mini-4k-instruct
1982
+ dtype: int64
1983
+ - name: mistralai/tekken
1984
+ dtype: int64
1985
+ - name: tiktoken/gpt-4o
1986
+ dtype: int64
1987
+ - name: tokenmonster/englishcode-32000-consistent-v1
1988
+ dtype: int64
1989
+ splits:
1990
+ - name: test
1991
+ num_bytes: 21709
1992
+ num_examples: 41
1993
+ download_size: 39820
1994
+ dataset_size: 21709
1995
  configs:
1996
  - config_name: tokenizer_robustness_completion_english_abbreviations
1997
  data_files:
 
2053
  data_files:
2054
  - split: test
2055
  path: tokenizer_robustness_completion_english_inflections/test-*
2056
+ - config_name: tokenizer_robustness_completion_english_keyboard_proximity_errors
2057
+ data_files:
2058
+ - split: test
2059
+ path: tokenizer_robustness_completion_english_keyboard_proximity_errors/test-*
2060
  ---
2061
 
2062
  # Dataset Card for Tokenization Robustness
tokenizer_robustness_completion_english_keyboard_proximity_errors/test-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:def8f5ea524b412225c00c8f0daf6d2694295ae083e495d1658b42cce9818fbc
3
+ size 39820