Update README.md

README.md

@@ -27,6 +27,9 @@ The model files can be used in [ComfyUI](https://github.com/comfyanonymous/Comfy
 
 ### Notes
 
-As you can see on the example above, Q3_K_M and Q3_K_S were pretty noisy, so I skipped them for now.
+> [!NOTE]
+> The Q5_K_M, Q4_K_M and most importantly the low bitrate quants (Q3_K_M, Q3_K_S, Q2_K) use a new dynamic logic where the first/last layer is kept in high precision.
+> 
+> For a comparison, see this [imgsli page](https://imgsli.com/NDA0MTIy). With this method, even Q2_K remains somewhat usable.
 
 *As this is a quantized model not a finetune, all the same restrictions/original license terms still apply.*