gsaltintas commited on
Commit
0659e92
·
verified ·
1 Parent(s): 239f215

Uploading tokenizer_robustness_completion_stem_latex subset

Browse files
README.md CHANGED
@@ -1124,6 +1124,130 @@ dataset_info:
1124
  num_examples: 17
1125
  download_size: 31032
1126
  dataset_size: 10163
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1127
  configs:
1128
  - config_name: tokenizer_robustness_completion_stem_canonical
1129
  data_files:
@@ -1161,6 +1285,10 @@ configs:
1161
  data_files:
1162
  - split: test
1163
  path: tokenizer_robustness_completion_stem_fullwidth_characters/test-*
 
 
 
 
1164
  ---
1165
 
1166
  # Dataset Card for Tokenization Robustness
 
1124
  num_examples: 17
1125
  download_size: 31032
1126
  dataset_size: 10163
1127
+ - config_name: tokenizer_robustness_completion_stem_latex
1128
+ features:
1129
+ - name: question
1130
+ dtype: string
1131
+ - name: choices
1132
+ list: string
1133
+ - name: answer
1134
+ dtype: int64
1135
+ - name: answer_label
1136
+ dtype: string
1137
+ - name: split
1138
+ dtype: string
1139
+ - name: subcategories
1140
+ dtype: string
1141
+ - name: lang
1142
+ dtype: string
1143
+ - name: second_lang
1144
+ dtype: string
1145
+ - name: notes
1146
+ dtype: string
1147
+ - name: id
1148
+ dtype: string
1149
+ - name: set_id
1150
+ dtype: string
1151
+ - name: variation_id
1152
+ dtype: string
1153
+ - name: question_general_category
1154
+ dtype: string
1155
+ - name: vanilla_cos_sim_to_canonical
1156
+ struct:
1157
+ - name: CohereLabs/aya-expanse-8b
1158
+ dtype: float64
1159
+ - name: Qwen/Qwen3-8B
1160
+ dtype: float64
1161
+ - name: bigscience/bloom
1162
+ dtype: float64
1163
+ - name: common-pile/comma-v0.1-1t
1164
+ dtype: float64
1165
+ - name: facebook/xglm-564M
1166
+ dtype: float64
1167
+ - name: google-bert/bert-base-multilingual-cased
1168
+ dtype: float64
1169
+ - name: google/byt5-small
1170
+ dtype: float64
1171
+ - name: google/gemma-2-2b
1172
+ dtype: float64
1173
+ - name: gpt2
1174
+ dtype: float64
1175
+ - name: meta-llama/Llama-3.2-1B
1176
+ dtype: float64
1177
+ - name: microsoft/Phi-3-mini-4k-instruct
1178
+ dtype: float64
1179
+ - name: mistralai/tekken
1180
+ dtype: float64
1181
+ - name: tiktoken/gpt-4o
1182
+ dtype: float64
1183
+ - name: tokenmonster/englishcode-32000-consistent-v1
1184
+ dtype: float64
1185
+ - name: trimmed_cos_sim_to_canonical
1186
+ struct:
1187
+ - name: CohereLabs/aya-expanse-8b
1188
+ dtype: float64
1189
+ - name: Qwen/Qwen3-8B
1190
+ dtype: float64
1191
+ - name: bigscience/bloom
1192
+ dtype: float64
1193
+ - name: common-pile/comma-v0.1-1t
1194
+ dtype: float64
1195
+ - name: facebook/xglm-564M
1196
+ dtype: float64
1197
+ - name: google-bert/bert-base-multilingual-cased
1198
+ dtype: float64
1199
+ - name: google/byt5-small
1200
+ dtype: float64
1201
+ - name: google/gemma-2-2b
1202
+ dtype: float64
1203
+ - name: gpt2
1204
+ dtype: float64
1205
+ - name: meta-llama/Llama-3.2-1B
1206
+ dtype: float64
1207
+ - name: microsoft/Phi-3-mini-4k-instruct
1208
+ dtype: float64
1209
+ - name: mistralai/tekken
1210
+ dtype: float64
1211
+ - name: tiktoken/gpt-4o
1212
+ dtype: float64
1213
+ - name: tokenmonster/englishcode-32000-consistent-v1
1214
+ dtype: float64
1215
+ - name: token_counts
1216
+ struct:
1217
+ - name: CohereLabs/aya-expanse-8b
1218
+ dtype: int64
1219
+ - name: Qwen/Qwen3-8B
1220
+ dtype: int64
1221
+ - name: bigscience/bloom
1222
+ dtype: int64
1223
+ - name: common-pile/comma-v0.1-1t
1224
+ dtype: int64
1225
+ - name: facebook/xglm-564M
1226
+ dtype: int64
1227
+ - name: google-bert/bert-base-multilingual-cased
1228
+ dtype: int64
1229
+ - name: google/byt5-small
1230
+ dtype: int64
1231
+ - name: google/gemma-2-2b
1232
+ dtype: int64
1233
+ - name: gpt2
1234
+ dtype: int64
1235
+ - name: meta-llama/Llama-3.2-1B
1236
+ dtype: int64
1237
+ - name: microsoft/Phi-3-mini-4k-instruct
1238
+ dtype: int64
1239
+ - name: mistralai/tekken
1240
+ dtype: int64
1241
+ - name: tiktoken/gpt-4o
1242
+ dtype: int64
1243
+ - name: tokenmonster/englishcode-32000-consistent-v1
1244
+ dtype: int64
1245
+ splits:
1246
+ - name: test
1247
+ num_bytes: 35181
1248
+ num_examples: 68
1249
+ download_size: 44974
1250
+ dataset_size: 35181
1251
  configs:
1252
  - config_name: tokenizer_robustness_completion_stem_canonical
1253
  data_files:
 
1285
  data_files:
1286
  - split: test
1287
  path: tokenizer_robustness_completion_stem_fullwidth_characters/test-*
1288
+ - config_name: tokenizer_robustness_completion_stem_latex
1289
+ data_files:
1290
+ - split: test
1291
+ path: tokenizer_robustness_completion_stem_latex/test-*
1292
  ---
1293
 
1294
  # Dataset Card for Tokenization Robustness
tokenizer_robustness_completion_stem_latex/test-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d5c5ffebfeaec4c598bbb56df699e0e74ca0fae4f07afda1f80b3262af2c8ca6
3
+ size 44974