Update README.md

README.md

@@ -45,7 +45,7 @@ The model implements several key improvements over standard RWKV architectures:
 
 ### Hybrid Design Benefits
 
-- **Linear Attention Inference**: RWKV blocks enable O(1) memory complexity during inference
+- **Linear Attention Inference**: RWKV blocks enable O(1) memory complexity during inference, and hybrids reduce memory usage by 1/7 KVCache.
 - **Enhanced Needle Tasks**: Strategic placement of GQA layers significantly improves performance on needle-in-haystack retrieval tasks, addressing a known limitation of pure linear attention models
 - **Implicit Position Encoding**: Interestingly, the model achieves better performance when RoPE (Rotary Position Embedding) is not applied to GQA layers, suggesting that RWKV blocks provide implicit positional encoding capabilities