Upload README.md with huggingface_hub

README.md

@@ -9,38 +9,19 @@ task_categories:
 pretty_name: Keep it simple !
 size_categories:
 - 100K<n<1M
-configs:
-- config_name: default
-  data_files:
-  - split: train
-    path: data/train-*
-dataset_info:
-  features:
-  - name: prompt
-    dtype: string
-  - name: text
-    dtype: string
-  - name: seed_data
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 914741887
-    num_examples: 640646
-  download_size: 513019825
-  dataset_size: 914741887
 ---
 
 
 # keep-it-simple
 
-**Objective:** An ultra-minimalist dataset for pre-training tiny language models. The logic relies on **bidirectional symmetry** (**A is B and B is A**) to foster deep semantic understanding. By training the model to predict the "prompt" from the "text" and vice versa, we maximize the utility of every pair.
+**Objective:** An ultra-minimalist dataset for pre-training tiny language models. The logic relies on **bidirectional symmetry** (**A is B and B is A]**) to foster deep semantic understanding. By training the model to predict the "prompt" from the "text" and vice versa, we maximize the utility of every pair.
 
 ## Data Sources
 
 * **Simple English Wikipedia**: Simplified encyclopedic articles.
 * **Vikidia (FR)**: Educational content for younger audiences.
-* **OPUS Books (en-fr)**: Aligned English-French literary translations.
-* **Cosmopedia-100k**: Synthetic educational content.
+* **[OPUS Books (en-fr)](https://aclanthology.org/L12-1246/)**: Aligned English-French literary translations.
+* **[Cosmopedia-100k](https://huggingface.co/blog/cosmopedia)**: Synthetic educational content.
 
 ## Structure
 
@@ -52,7 +33,7 @@ dataset_info:
 
 ## Context & Usage
 
-* **Bidirectional Training**: Each source item yields two training entries (`prompt` $\rightarrow$ `text` and `text` $\rightarrow$ `prompt`). This enforces semantic symmetry.
+* **Bidirectional Training**: Each source item yields two training entries (`prompt` $\rightarrow$ `text` and `text` $\rightarrow$ `prompt`). This enforces semantic symmetry, [reversal curve](https://arxiv.org/abs/2309.12288) and limitate span corruption.
 * **Minimalism**: More compact than the **[BabyLM](https://arxiv.org/abs/2602.20092)** challenge; focused on density and the purity of pairs to maximize efficiency on tiny, resource-constrained architectures.
 * **Goal**: Rapid testing of alignment theories and training "pocket" models for fundamental, bidirectional interactions.