gsaltintas commited on
Commit
2d2d1c8
·
verified ·
1 Parent(s): e0ab80c

Uploading tokenizer_robustness_completion_english_superscript_subscript_styling subset

Browse files
README.md CHANGED
@@ -3232,6 +3232,130 @@ dataset_info:
3232
  num_examples: 15
3233
  download_size: 30689
3234
  dataset_size: 8091
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3235
  configs:
3236
  - config_name: tokenizer_robustness_completion_english_abbreviations
3237
  data_files:
@@ -3337,6 +3461,10 @@ configs:
3337
  data_files:
3338
  - split: test
3339
  path: tokenizer_robustness_completion_english_spelled_out/test-*
 
 
 
 
3340
  ---
3341
 
3342
  # Dataset Card for Tokenization Robustness
 
3232
  num_examples: 15
3233
  download_size: 30689
3234
  dataset_size: 8091
3235
+ - config_name: tokenizer_robustness_completion_english_superscript_subscript_styling
3236
+ features:
3237
+ - name: question
3238
+ dtype: string
3239
+ - name: choices
3240
+ list: string
3241
+ - name: answer
3242
+ dtype: int64
3243
+ - name: answer_label
3244
+ dtype: string
3245
+ - name: split
3246
+ dtype: string
3247
+ - name: subcategories
3248
+ dtype: string
3249
+ - name: category
3250
+ dtype: string
3251
+ - name: lang
3252
+ dtype: string
3253
+ - name: second_lang
3254
+ dtype: string
3255
+ - name: notes
3256
+ dtype: string
3257
+ - name: id
3258
+ dtype: string
3259
+ - name: set_id
3260
+ dtype: string
3261
+ - name: variation_id
3262
+ dtype: string
3263
+ - name: vanilla_cos_sim_to_canonical
3264
+ struct:
3265
+ - name: CohereLabs/aya-expanse-8b
3266
+ dtype: float64
3267
+ - name: Qwen/Qwen3-8B
3268
+ dtype: float64
3269
+ - name: bigscience/bloom
3270
+ dtype: float64
3271
+ - name: common-pile/comma-v0.1-1t
3272
+ dtype: float64
3273
+ - name: facebook/xglm-564M
3274
+ dtype: float64
3275
+ - name: google-bert/bert-base-multilingual-cased
3276
+ dtype: float64
3277
+ - name: google/byt5-small
3278
+ dtype: float64
3279
+ - name: google/gemma-2-2b
3280
+ dtype: float64
3281
+ - name: gpt2
3282
+ dtype: float64
3283
+ - name: meta-llama/Llama-3.2-1B
3284
+ dtype: float64
3285
+ - name: microsoft/Phi-3-mini-4k-instruct
3286
+ dtype: float64
3287
+ - name: mistralai/tekken
3288
+ dtype: float64
3289
+ - name: tiktoken/gpt-4o
3290
+ dtype: float64
3291
+ - name: tokenmonster/englishcode-32000-consistent-v1
3292
+ dtype: float64
3293
+ - name: trimmed_cos_sim_to_canonical
3294
+ struct:
3295
+ - name: CohereLabs/aya-expanse-8b
3296
+ dtype: float64
3297
+ - name: Qwen/Qwen3-8B
3298
+ dtype: float64
3299
+ - name: bigscience/bloom
3300
+ dtype: float64
3301
+ - name: common-pile/comma-v0.1-1t
3302
+ dtype: float64
3303
+ - name: facebook/xglm-564M
3304
+ dtype: float64
3305
+ - name: google-bert/bert-base-multilingual-cased
3306
+ dtype: float64
3307
+ - name: google/byt5-small
3308
+ dtype: float64
3309
+ - name: google/gemma-2-2b
3310
+ dtype: float64
3311
+ - name: gpt2
3312
+ dtype: float64
3313
+ - name: meta-llama/Llama-3.2-1B
3314
+ dtype: float64
3315
+ - name: microsoft/Phi-3-mini-4k-instruct
3316
+ dtype: float64
3317
+ - name: mistralai/tekken
3318
+ dtype: float64
3319
+ - name: tiktoken/gpt-4o
3320
+ dtype: float64
3321
+ - name: tokenmonster/englishcode-32000-consistent-v1
3322
+ dtype: float64
3323
+ - name: token_counts
3324
+ struct:
3325
+ - name: CohereLabs/aya-expanse-8b
3326
+ dtype: int64
3327
+ - name: Qwen/Qwen3-8B
3328
+ dtype: int64
3329
+ - name: bigscience/bloom
3330
+ dtype: int64
3331
+ - name: common-pile/comma-v0.1-1t
3332
+ dtype: int64
3333
+ - name: facebook/xglm-564M
3334
+ dtype: int64
3335
+ - name: google-bert/bert-base-multilingual-cased
3336
+ dtype: int64
3337
+ - name: google/byt5-small
3338
+ dtype: int64
3339
+ - name: google/gemma-2-2b
3340
+ dtype: int64
3341
+ - name: gpt2
3342
+ dtype: int64
3343
+ - name: meta-llama/Llama-3.2-1B
3344
+ dtype: int64
3345
+ - name: microsoft/Phi-3-mini-4k-instruct
3346
+ dtype: int64
3347
+ - name: mistralai/tekken
3348
+ dtype: int64
3349
+ - name: tiktoken/gpt-4o
3350
+ dtype: int64
3351
+ - name: tokenmonster/englishcode-32000-consistent-v1
3352
+ dtype: int64
3353
+ splits:
3354
+ - name: test
3355
+ num_bytes: 25705
3356
+ num_examples: 40
3357
+ download_size: 41912
3358
+ dataset_size: 25705
3359
  configs:
3360
  - config_name: tokenizer_robustness_completion_english_abbreviations
3361
  data_files:
 
3461
  data_files:
3462
  - split: test
3463
  path: tokenizer_robustness_completion_english_spelled_out/test-*
3464
+ - config_name: tokenizer_robustness_completion_english_superscript_subscript_styling
3465
+ data_files:
3466
+ - split: test
3467
+ path: tokenizer_robustness_completion_english_superscript_subscript_styling/test-*
3468
  ---
3469
 
3470
  # Dataset Card for Tokenization Robustness
tokenizer_robustness_completion_english_superscript_subscript_styling/test-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:816af9becd6f37f7b117aad4461bd83c1a2dc8dc8684ef5afc62c6d7287d95d4
3
+ size 41912