gsaltintas commited on
Commit
a1fb36a
·
verified ·
1 Parent(s): 44b6ea3

Uploading tokenizer_robustness_completion_farsi_equivalent_expressions subset

Browse files
README.md CHANGED
@@ -28,6 +28,10 @@ configs:
28
  data_files:
29
  - split: test
30
  path: tokenizer_robustness_completion_farsi_dialects/test-*
 
 
 
 
31
  dataset_info:
32
  - config_name: tokenizer_robustness_completion_farsi_arabic_keyboard_for_farsi
33
  features:
@@ -649,6 +653,130 @@ dataset_info:
649
  num_examples: 360
650
  download_size: 105228
651
  dataset_size: 217454
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
652
  ---
653
 
654
  # Dataset Card for Tokenization Robustness
 
28
  data_files:
29
  - split: test
30
  path: tokenizer_robustness_completion_farsi_dialects/test-*
31
+ - config_name: tokenizer_robustness_completion_farsi_equivalent_expressions
32
+ data_files:
33
+ - split: test
34
+ path: tokenizer_robustness_completion_farsi_equivalent_expressions/test-*
35
  dataset_info:
36
  - config_name: tokenizer_robustness_completion_farsi_arabic_keyboard_for_farsi
37
  features:
 
653
  num_examples: 360
654
  download_size: 105228
655
  dataset_size: 217454
656
+ - config_name: tokenizer_robustness_completion_farsi_equivalent_expressions
657
+ features:
658
+ - name: question
659
+ dtype: string
660
+ - name: choices
661
+ list: string
662
+ - name: answer
663
+ dtype: int64
664
+ - name: answer_label
665
+ dtype: string
666
+ - name: split
667
+ dtype: string
668
+ - name: subcategories
669
+ dtype: string
670
+ - name: category
671
+ dtype: string
672
+ - name: lang
673
+ dtype: string
674
+ - name: second_lang
675
+ dtype: string
676
+ - name: notes
677
+ dtype: string
678
+ - name: id
679
+ dtype: string
680
+ - name: set_id
681
+ dtype: string
682
+ - name: variation_id
683
+ dtype: string
684
+ - name: vanilla_cos_sim_to_canonical
685
+ struct:
686
+ - name: CohereLabs/aya-expanse-8b
687
+ dtype: float64
688
+ - name: Qwen/Qwen3-8B
689
+ dtype: float64
690
+ - name: bigscience/bloom
691
+ dtype: float64
692
+ - name: common-pile/comma-v0.1-1t
693
+ dtype: float64
694
+ - name: facebook/xglm-564M
695
+ dtype: float64
696
+ - name: google-bert/bert-base-multilingual-cased
697
+ dtype: float64
698
+ - name: google/byt5-small
699
+ dtype: float64
700
+ - name: google/gemma-2-2b
701
+ dtype: float64
702
+ - name: gpt2
703
+ dtype: float64
704
+ - name: meta-llama/Llama-3.2-1B
705
+ dtype: float64
706
+ - name: microsoft/Phi-3-mini-4k-instruct
707
+ dtype: float64
708
+ - name: mistralai/tekken
709
+ dtype: float64
710
+ - name: tiktoken/gpt-4o
711
+ dtype: float64
712
+ - name: tokenmonster/englishcode-32000-consistent-v1
713
+ dtype: float64
714
+ - name: trimmed_cos_sim_to_canonical
715
+ struct:
716
+ - name: CohereLabs/aya-expanse-8b
717
+ dtype: float64
718
+ - name: Qwen/Qwen3-8B
719
+ dtype: float64
720
+ - name: bigscience/bloom
721
+ dtype: float64
722
+ - name: common-pile/comma-v0.1-1t
723
+ dtype: float64
724
+ - name: facebook/xglm-564M
725
+ dtype: float64
726
+ - name: google-bert/bert-base-multilingual-cased
727
+ dtype: float64
728
+ - name: google/byt5-small
729
+ dtype: float64
730
+ - name: google/gemma-2-2b
731
+ dtype: float64
732
+ - name: gpt2
733
+ dtype: float64
734
+ - name: meta-llama/Llama-3.2-1B
735
+ dtype: float64
736
+ - name: microsoft/Phi-3-mini-4k-instruct
737
+ dtype: float64
738
+ - name: mistralai/tekken
739
+ dtype: float64
740
+ - name: tiktoken/gpt-4o
741
+ dtype: float64
742
+ - name: tokenmonster/englishcode-32000-consistent-v1
743
+ dtype: float64
744
+ - name: token_counts
745
+ struct:
746
+ - name: CohereLabs/aya-expanse-8b
747
+ dtype: int64
748
+ - name: Qwen/Qwen3-8B
749
+ dtype: int64
750
+ - name: bigscience/bloom
751
+ dtype: int64
752
+ - name: common-pile/comma-v0.1-1t
753
+ dtype: int64
754
+ - name: facebook/xglm-564M
755
+ dtype: int64
756
+ - name: google-bert/bert-base-multilingual-cased
757
+ dtype: int64
758
+ - name: google/byt5-small
759
+ dtype: int64
760
+ - name: google/gemma-2-2b
761
+ dtype: int64
762
+ - name: gpt2
763
+ dtype: int64
764
+ - name: meta-llama/Llama-3.2-1B
765
+ dtype: int64
766
+ - name: microsoft/Phi-3-mini-4k-instruct
767
+ dtype: int64
768
+ - name: mistralai/tekken
769
+ dtype: int64
770
+ - name: tiktoken/gpt-4o
771
+ dtype: int64
772
+ - name: tokenmonster/englishcode-32000-consistent-v1
773
+ dtype: int64
774
+ splits:
775
+ - name: test
776
+ num_bytes: 24339
777
+ num_examples: 40
778
+ download_size: 41138
779
+ dataset_size: 24339
780
  ---
781
 
782
  # Dataset Card for Tokenization Robustness
tokenizer_robustness_completion_farsi_equivalent_expressions/test-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9ce0ca75dfd0937903e7994c274a99e5b72f266fee41451c7e41e494afd0fe97
3
+ size 41138