Update README.md

README.md

@@ -28,13 +28,13 @@ tags:
 
 This is quantized version of the original uncensored model Dolphin 3.0 Qwen 2.5 1.5B.
 
-The term "uncensured" or "unrestricted censoring," in a context related to AI models like this Dolphin's Uncensored Model, typically refers to the absence of any form of content control, filtering, editing tools during training and development. This means that all possible inputs are allowed without restriction.
+The term "uncensored" or "unrestricted censoring," in a context related to AI models like this Dolphin's Uncensored Model, typically refers to the absence of any form of content control, filtering, editing tools during training and development. This means that all possible inputs are allowed without restriction.
 
 The importance lies primarily on two aspects:
 
 1) **Content Generation**: Without censorship or restrictions in place for data collection from various sources (like social media platforms where users could post anything), the model would have access to a broader range of content, including potentially sensitive information. This can lead to more diverse and comprehensive training datasets.
 
-2) **Adaptability & Generalization:** A un-censored approach allows models like DQW-UNC Dolphin 3.0 Qwen with large language capacity (1B tokens), such as the recently released version known for its impressive performance, including a significant leap in understanding complex human conversations and context-aware responses.
+2) **Adaptability & Generalization:** A un-censored approach allows models like Dolphin 3.0 Qwen with large language capacity (1B tokens), such as the recently released version known for its impressive performance, including a significant leap in understanding complex human conversations and context-aware responses.
 
 However it's important to note that while uncensured can lead to more diverse data sources which could potentially improve model generalization capabilities through exposure of various linguistic patterns or cultural nuances not typically seen within restricted datasets.