inkoziev commited on
Commit
b7f7b60
1 Parent(s): 0425c6a

The description of the error typology has been brought into line with the actual contents of the error_type field

Browse files
Files changed (1) hide show
  1. README.md +63 -10
README.md CHANGED
@@ -34,7 +34,7 @@ In addition, for some samples there is a corrected version of the text ("fixed_s
34
  "sentence": "Разучи стихов по больше",
35
  "fixed_sentence": "Разучи стихов побольше",
36
  "label": 0,
37
- "error_type": "Misspelling",
38
  "domain": "prose"
39
  }
40
  ```
@@ -64,7 +64,7 @@ A few poetry samples are included in this version:
64
  "sentence": "Чему научит забытьё?\nСмерть формы д'арует литьё.\nРезец мгновенье любит стружка...\nСмерть безобидная подружка!",
65
  "fixed_sentence": null,
66
  "label": 0,
67
- "error_type": "Syntax",
68
  "domain": "poetry"
69
  }
70
  ```
@@ -77,22 +77,75 @@ A few poetry samples are included in this version:
77
  **sentence** (str): the original sentence.
78
  **fixed_sentence** (str): the corrected version of original sentence.
79
  **label** (str): the target class. "1" for "acceptable", "0" for "unacceptable".
80
- **error_type** (str): the violation category: Misspelling, Syntax, Semantics, or Hallucination.
81
  **domain** (str): domain: "prose" or "poetry".
82
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
83
  ### Statistics
84
 
85
 
86
- Total number of samples in test split: **5829**
87
  Total number of samples in train split: **435538**
88
 
89
  Statistics for test split.
90
 
91
  Domains:
92
- prose 5220
93
  poetry 609
94
-
95
- Violation categories:
96
- Misspelling 3266
97
- Syntax 2555
98
- Semantics 8
 
34
  "sentence": "Разучи стихов по больше",
35
  "fixed_sentence": "Разучи стихов побольше",
36
  "label": 0,
37
+ "error_type": "Tokenization",
38
  "domain": "prose"
39
  }
40
  ```
 
64
  "sentence": "Чему научит забытьё?\nСмерть формы д'арует литьё.\nРезец мгновенье любит стружка...\nСмерть безобидная подружка!",
65
  "fixed_sentence": null,
66
  "label": 0,
67
+ "error_type": "Grammar",
68
  "domain": "poetry"
69
  }
70
  ```
 
77
  **sentence** (str): the original sentence.
78
  **fixed_sentence** (str): the corrected version of original sentence.
79
  **label** (str): the target class. "1" for "acceptable", "0" for "unacceptable".
80
+ **error_type** (str): the violation category: Spelling, Grammar, Tokenization, Punctuation.
81
  **domain** (str): domain: "prose" or "poetry".
82
 
83
+ ### Error types
84
+
85
+ **Tokenization**: a word is split into two tokens, or two words are merged into one word.
86
+
87
+ ```
88
+ {
89
+ "id": 6,
90
+ "sentence": "Я подбираю по проще слова",
91
+ "fixed_sentence": "Я подбираю попроще слова",
92
+ "label": 0,
93
+ "error_type": "Tokenization",
94
+ "domain": "prose"
95
+ }
96
+ ```
97
+
98
+ **Punctuation**: missing or extra comma, hyphen or other punctuation mark
99
+
100
+ ```
101
+ {
102
+ "id": 5,
103
+ "sentence": "И швырнуть по-дальше",
104
+ "fixed_sentence": "И швырнуть подальше",
105
+ "label": 0,
106
+ "error_type": "Punctuation",
107
+ "domain": "prose"
108
+ }
109
+ ```
110
+
111
+ **Spelling**:
112
+
113
+ ```
114
+ {
115
+ "id": 38,
116
+ "sentence": "И ведь что интересно, русские официально ни в одном крестовом позоде не участвовали.",
117
+ "fixed_sentence": "И ведь что интересно, русские официально ни в одном крестовом походе не участвовали.",
118
+ "label": 0,
119
+ "error_type": "Spelling",
120
+ "domain": "prose"
121
+ }
122
+ ```
123
+
124
+ **Grammar**: One of the words is in the wrong grammatical form, for example the verb is in the infinitive instead of the personal form.
125
+
126
+ ```
127
+ {
128
+ "id": 61,
129
+ "sentence": "на него никто не польститься",
130
+ "fixed_sentence": "на него никто не польстится",
131
+ "label": 0,
132
+ "error_type": "Grammar",
133
+ "domain": "prose"
134
+ }
135
+ ```
136
+
137
+ Please note that error categories are not always set accurately, so you should not use
138
+ the "error_type" field to train classifiers.
139
+
140
+
141
  ### Statistics
142
 
143
 
144
+ Total number of samples in test split: **6244**
145
  Total number of samples in train split: **435538**
146
 
147
  Statistics for test split.
148
 
149
  Domains:
150
+ prose 5635
151
  poetry 609