Upload README.md with huggingface_hub

README.md

@@ -49,11 +49,6 @@ Each `.jsonl` file contains entries with the following fields:
 **Thresholds tested:** `0.5 – 0.8`  
 **Final dataset:** `/parallel_data/matched_0.60_0.005.jsonl` ("best" match)
 
-### Dataset Quality Note
-
-The dataset was aligned using a relatively low cosine similarity threshold (e.g. 0.60) to increase the number of matches. However, this likely reduces alignment quality and may introduce noisy or weakly related sentence pairs. Further filtering or manual inspection would be necessary to improve reliability.
-
-
 ## Threshold and Alpha Comparison
 
 We compared cosine similarity thresholds (0.50–0.80) using two alpha settings: alpha=None (no length regularization) and alpha=0.005 (penalizing length mismatch). alpha=None produced more matches, especially at low thresholds, but with a steep quality drop as the threshold increased. alpha=0.005 yielded fewer but more consistent matches.
@@ -61,6 +56,10 @@ We compared cosine similarity thresholds (0.50–0.80) using two alpha settings:
 Conclusion:
 For better balance between match quantity and alignment quality, we selected alpha=0.005 with a threshold of 0.60`.
 
+### Dataset Quality Note
+
+As the dataset was aligned using a relatively low cosine similarity threshold (e.g. 0.60) to increase the number of matches it likely reduces alignment quality and may introduce noisy or weakly related sentence pairs. Further filtering or manual inspection would be necessary to improve reliability.
+
 
 ## Evaluation
 
@@ -78,4 +77,5 @@ Unit tests confirmed:
 | --------------------- | ---------------------------------- |
 | `parallel_data/`  | All alignment outputs by threshold |
 | `/parallel_data/matched_0.60_0.005.jsonl` | "best match"     |
-| `embedded`           | German and Romansh embeddings                       |
+| `embedded`           | German and Romansh embeddings  |
+| `data/`               | Original data