milwright commited on
Commit
54cf274
·
1 Parent(s): 01a1595

add hugging face space metadata to readme

Browse files
Files changed (1) hide show
  1. README.md +9 -2
README.md CHANGED
@@ -1,4 +1,11 @@
1
- # Cloze Reader
 
 
 
 
 
 
 
2
 
3
  An interactive reading comprehension game using AI to generate cloze (fill-in-the-blank) exercises from public domain literature.
4
 
@@ -10,7 +17,7 @@ The Cloze Reader transforms passages from Project Gutenberg into adaptive vocabu
10
 
11
  **Educational cloze testing (1953)**: Wilson L. Taylor introduced the cloze procedure—systematically deleting words from passages to measure reading comprehension. It became standard in U.S. educational assessment by the 1960s.
12
 
13
- **Masked language modeling (2018)**: BERT and subsequent models rediscovered cloze methodology independently as a training objective, randomly masking tokens and predicting from context. Both domains converge on the same principle: understanding language means predicting from context.
14
 
15
  **This project**: Uses language models trained on prediction tasks to generate prediction exercises for human readers. While Gemma-3 uses next-token prediction rather than masked language modeling, the system demonstrates how assessment and training methodologies are now instrumentalized through identical computational systems.
16
 
 
1
+ ---
2
+ title: Cloze Reader
3
+ emoji: 📚
4
+ colorFrom: yellow
5
+ colorTo: gray
6
+ sdk: docker
7
+ pinned: true
8
+ ---
9
 
10
  An interactive reading comprehension game using AI to generate cloze (fill-in-the-blank) exercises from public domain literature.
11
 
 
17
 
18
  **Educational cloze testing (1953)**: Wilson L. Taylor introduced the cloze procedure—systematically deleting words from passages to measure reading comprehension. It became standard in U.S. educational assessment by the 1960s.
19
 
20
+ **Masked language modeling (2018)**: BERT and subsequent models rediscovered cloze methodology independently as a training objective, randomly masking tokens and predicting from context.
21
 
22
  **This project**: Uses language models trained on prediction tasks to generate prediction exercises for human readers. While Gemma-3 uses next-token prediction rather than masked language modeling, the system demonstrates how assessment and training methodologies are now instrumentalized through identical computational systems.
23