gsaltintas commited on
Commit
be62875
·
verified ·
1 Parent(s): fdee5ec

Uploading tokenizer_robustness_completion_english_scripted_text subset

Browse files
README.md CHANGED
@@ -2612,6 +2612,130 @@ dataset_info:
2612
  num_examples: 80
2613
  download_size: 49627
2614
  dataset_size: 43643
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2615
  configs:
2616
  - config_name: tokenizer_robustness_completion_english_abbreviations
2617
  data_files:
@@ -2697,6 +2821,10 @@ configs:
2697
  data_files:
2698
  - split: test
2699
  path: tokenizer_robustness_completion_english_orthographic_errors/test-*
 
 
 
 
2700
  ---
2701
 
2702
  # Dataset Card for Tokenization Robustness
 
2612
  num_examples: 80
2613
  download_size: 49627
2614
  dataset_size: 43643
2615
+ - config_name: tokenizer_robustness_completion_english_scripted_text
2616
+ features:
2617
+ - name: question
2618
+ dtype: string
2619
+ - name: choices
2620
+ list: string
2621
+ - name: answer
2622
+ dtype: int64
2623
+ - name: answer_label
2624
+ dtype: string
2625
+ - name: split
2626
+ dtype: string
2627
+ - name: subcategories
2628
+ dtype: string
2629
+ - name: category
2630
+ dtype: string
2631
+ - name: lang
2632
+ dtype: string
2633
+ - name: second_lang
2634
+ dtype: string
2635
+ - name: notes
2636
+ dtype: string
2637
+ - name: id
2638
+ dtype: string
2639
+ - name: set_id
2640
+ dtype: string
2641
+ - name: variation_id
2642
+ dtype: string
2643
+ - name: vanilla_cos_sim_to_canonical
2644
+ struct:
2645
+ - name: CohereLabs/aya-expanse-8b
2646
+ dtype: float64
2647
+ - name: Qwen/Qwen3-8B
2648
+ dtype: float64
2649
+ - name: bigscience/bloom
2650
+ dtype: float64
2651
+ - name: common-pile/comma-v0.1-1t
2652
+ dtype: float64
2653
+ - name: facebook/xglm-564M
2654
+ dtype: float64
2655
+ - name: google-bert/bert-base-multilingual-cased
2656
+ dtype: float64
2657
+ - name: google/byt5-small
2658
+ dtype: float64
2659
+ - name: google/gemma-2-2b
2660
+ dtype: float64
2661
+ - name: gpt2
2662
+ dtype: float64
2663
+ - name: meta-llama/Llama-3.2-1B
2664
+ dtype: float64
2665
+ - name: microsoft/Phi-3-mini-4k-instruct
2666
+ dtype: float64
2667
+ - name: mistralai/tekken
2668
+ dtype: float64
2669
+ - name: tiktoken/gpt-4o
2670
+ dtype: float64
2671
+ - name: tokenmonster/englishcode-32000-consistent-v1
2672
+ dtype: float64
2673
+ - name: trimmed_cos_sim_to_canonical
2674
+ struct:
2675
+ - name: CohereLabs/aya-expanse-8b
2676
+ dtype: float64
2677
+ - name: Qwen/Qwen3-8B
2678
+ dtype: float64
2679
+ - name: bigscience/bloom
2680
+ dtype: float64
2681
+ - name: common-pile/comma-v0.1-1t
2682
+ dtype: float64
2683
+ - name: facebook/xglm-564M
2684
+ dtype: float64
2685
+ - name: google-bert/bert-base-multilingual-cased
2686
+ dtype: float64
2687
+ - name: google/byt5-small
2688
+ dtype: float64
2689
+ - name: google/gemma-2-2b
2690
+ dtype: float64
2691
+ - name: gpt2
2692
+ dtype: float64
2693
+ - name: meta-llama/Llama-3.2-1B
2694
+ dtype: float64
2695
+ - name: microsoft/Phi-3-mini-4k-instruct
2696
+ dtype: float64
2697
+ - name: mistralai/tekken
2698
+ dtype: float64
2699
+ - name: tiktoken/gpt-4o
2700
+ dtype: float64
2701
+ - name: tokenmonster/englishcode-32000-consistent-v1
2702
+ dtype: float64
2703
+ - name: token_counts
2704
+ struct:
2705
+ - name: CohereLabs/aya-expanse-8b
2706
+ dtype: int64
2707
+ - name: Qwen/Qwen3-8B
2708
+ dtype: int64
2709
+ - name: bigscience/bloom
2710
+ dtype: int64
2711
+ - name: common-pile/comma-v0.1-1t
2712
+ dtype: int64
2713
+ - name: facebook/xglm-564M
2714
+ dtype: int64
2715
+ - name: google-bert/bert-base-multilingual-cased
2716
+ dtype: int64
2717
+ - name: google/byt5-small
2718
+ dtype: int64
2719
+ - name: google/gemma-2-2b
2720
+ dtype: int64
2721
+ - name: gpt2
2722
+ dtype: int64
2723
+ - name: meta-llama/Llama-3.2-1B
2724
+ dtype: int64
2725
+ - name: microsoft/Phi-3-mini-4k-instruct
2726
+ dtype: int64
2727
+ - name: mistralai/tekken
2728
+ dtype: int64
2729
+ - name: tiktoken/gpt-4o
2730
+ dtype: int64
2731
+ - name: tokenmonster/englishcode-32000-consistent-v1
2732
+ dtype: int64
2733
+ splits:
2734
+ - name: test
2735
+ num_bytes: 27700
2736
+ num_examples: 40
2737
+ download_size: 42142
2738
+ dataset_size: 27700
2739
  configs:
2740
  - config_name: tokenizer_robustness_completion_english_abbreviations
2741
  data_files:
 
2821
  data_files:
2822
  - split: test
2823
  path: tokenizer_robustness_completion_english_orthographic_errors/test-*
2824
+ - config_name: tokenizer_robustness_completion_english_scripted_text
2825
+ data_files:
2826
+ - split: test
2827
+ path: tokenizer_robustness_completion_english_scripted_text/test-*
2828
  ---
2829
 
2830
  # Dataset Card for Tokenization Robustness
tokenizer_robustness_completion_english_scripted_text/test-00000-of-00001.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ca72aea22875e7498a670b1ebc5aaa7affaa0068992668e1ddd928adf9923974
3
+ size 42142