gsaltintas commited on
Commit
0b5f8fb
·
verified ·
1 Parent(s): d9e9232

Uploading tokenizer_robustness_completion_english_word_reordering subset

Browse files
README.md CHANGED
@@ -3480,6 +3480,130 @@ dataset_info:
3480
  num_examples: 37
3481
  download_size: 38294
3482
  dataset_size: 18991
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3483
  configs:
3484
  - config_name: tokenizer_robustness_completion_english_abbreviations
3485
  data_files:
@@ -3593,6 +3717,10 @@ configs:
3593
  data_files:
3594
  - split: test
3595
  path: tokenizer_robustness_completion_english_web_search_query/test-*
 
 
 
 
3596
  ---
3597
 
3598
  # Dataset Card for Tokenization Robustness
 
3480
  num_examples: 37
3481
  download_size: 38294
3482
  dataset_size: 18991
3483
+ - config_name: tokenizer_robustness_completion_english_word_reordering
3484
+ features:
3485
+ - name: question
3486
+ dtype: string
3487
+ - name: choices
3488
+ list: string
3489
+ - name: answer
3490
+ dtype: int64
3491
+ - name: answer_label
3492
+ dtype: string
3493
+ - name: split
3494
+ dtype: string
3495
+ - name: subcategories
3496
+ dtype: string
3497
+ - name: category
3498
+ dtype: string
3499
+ - name: lang
3500
+ dtype: string
3501
+ - name: second_lang
3502
+ dtype: string
3503
+ - name: notes
3504
+ dtype: string
3505
+ - name: id
3506
+ dtype: string
3507
+ - name: set_id
3508
+ dtype: string
3509
+ - name: variation_id
3510
+ dtype: string
3511
+ - name: vanilla_cos_sim_to_canonical
3512
+ struct:
3513
+ - name: CohereLabs/aya-expanse-8b
3514
+ dtype: float64
3515
+ - name: Qwen/Qwen3-8B
3516
+ dtype: float64
3517
+ - name: bigscience/bloom
3518
+ dtype: float64
3519
+ - name: common-pile/comma-v0.1-1t
3520
+ dtype: float64
3521
+ - name: facebook/xglm-564M
3522
+ dtype: float64
3523
+ - name: google-bert/bert-base-multilingual-cased
3524
+ dtype: float64
3525
+ - name: google/byt5-small
3526
+ dtype: float64
3527
+ - name: google/gemma-2-2b
3528
+ dtype: float64
3529
+ - name: gpt2
3530
+ dtype: float64
3531
+ - name: meta-llama/Llama-3.2-1B
3532
+ dtype: float64
3533
+ - name: microsoft/Phi-3-mini-4k-instruct
3534
+ dtype: float64
3535
+ - name: mistralai/tekken
3536
+ dtype: float64
3537
+ - name: tiktoken/gpt-4o
3538
+ dtype: float64
3539
+ - name: tokenmonster/englishcode-32000-consistent-v1
3540
+ dtype: float64
3541
+ - name: trimmed_cos_sim_to_canonical
3542
+ struct:
3543
+ - name: CohereLabs/aya-expanse-8b
3544
+ dtype: float64
3545
+ - name: Qwen/Qwen3-8B
3546
+ dtype: float64
3547
+ - name: bigscience/bloom
3548
+ dtype: float64
3549
+ - name: common-pile/comma-v0.1-1t
3550
+ dtype: float64
3551
+ - name: facebook/xglm-564M
3552
+ dtype: float64
3553
+ - name: google-bert/bert-base-multilingual-cased
3554
+ dtype: float64
3555
+ - name: google/byt5-small
3556
+ dtype: float64
3557
+ - name: google/gemma-2-2b
3558
+ dtype: float64
3559
+ - name: gpt2
3560
+ dtype: float64
3561
+ - name: meta-llama/Llama-3.2-1B
3562
+ dtype: float64
3563
+ - name: microsoft/Phi-3-mini-4k-instruct
3564
+ dtype: float64
3565
+ - name: mistralai/tekken
3566
+ dtype: float64
3567
+ - name: tiktoken/gpt-4o
3568
+ dtype: float64
3569
+ - name: tokenmonster/englishcode-32000-consistent-v1
3570
+ dtype: float64
3571
+ - name: token_counts
3572
+ struct:
3573
+ - name: CohereLabs/aya-expanse-8b
3574
+ dtype: int64
3575
+ - name: Qwen/Qwen3-8B
3576
+ dtype: int64
3577
+ - name: bigscience/bloom
3578
+ dtype: int64
3579
+ - name: common-pile/comma-v0.1-1t
3580
+ dtype: int64
3581
+ - name: facebook/xglm-564M
3582
+ dtype: int64
3583
+ - name: google-bert/bert-base-multilingual-cased
3584
+ dtype: int64
3585
+ - name: google/byt5-small
3586
+ dtype: int64
3587
+ - name: google/gemma-2-2b
3588
+ dtype: int64
3589
+ - name: gpt2
3590
+ dtype: int64
3591
+ - name: meta-llama/Llama-3.2-1B
3592
+ dtype: int64
3593
+ - name: microsoft/Phi-3-mini-4k-instruct
3594
+ dtype: int64
3595
+ - name: mistralai/tekken
3596
+ dtype: int64
3597
+ - name: tiktoken/gpt-4o
3598
+ dtype: int64
3599
+ - name: tokenmonster/englishcode-32000-consistent-v1
3600
+ dtype: int64
3601
+ splits:
3602
+ - name: test
3603
+ num_bytes: 21498
3604
+ num_examples: 40
3605
+ download_size: 39384
3606
+ dataset_size: 21498
3607
  configs:
3608
  - config_name: tokenizer_robustness_completion_english_abbreviations
3609
  data_files:
 
3717
  data_files:
3718
  - split: test
3719
  path: tokenizer_robustness_completion_english_web_search_query/test-*
3720
+ - config_name: tokenizer_robustness_completion_english_word_reordering
3721
+ data_files:
3722
+ - split: test
3723
+ path: tokenizer_robustness_completion_english_word_reordering/test-*
3724
  ---
3725
 
3726
  # Dataset Card for Tokenization Robustness
tokenizer_robustness_completion_english_word_reordering/test-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b35d2f928e6e12dcbc04cf54b9950405c6cda15e56ea99b33996f7d37df61dd5
3
+ size 39384