Update README.md

README.md

@@ -66,13 +66,13 @@ This makes BLOOM a representative example of multilingual BPE tokenization.
 
 ## Model Architecture
 
-- **Architecture:** Decoder-only Transformer (LLaMA-style)
+- **Architecture:**  Decoder-only Transformer (Lingua's Llama-3.2-1B configuration)
 - **Non-embedding parameters:** ~1B
 - **Context length:** 4096 tokens
 - **Framework:** Meta Lingua
-- **Initialization:** Shared super-vocabulary initialization for overlapping token strings
+- **Initialization:** Shared super-vocabulary initialization across TokSuite models
 
-The architecture and hyperparameters are fixed across TokSuite; the tokenizer is the only variable.
+The architecture and training setup are identical across all TokSuite models; only the tokenizer differs.
 
 ---