Update README.md

README.md

@@ -59,13 +59,15 @@ The training used 5*A800 for 3 epochs
   train_steps_per_second   =              0.224
 ```
 
-We can see from some preliminary results, the conversation is natural and informative (unsurprisingly), also we observe the unfiltering seems to be working! 
+We can see from some preliminary results, the conversation is natural and informative (unsurprisingly).
 
-**Heads up** some examples are unsafe and inappropriate, this is entirely for the purpose of testing how un-aligned SFT data affect LLM's final output.
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/6413d7be996b2e426f230fb7/WfQYyyLxtXA2KlePmIPQJ.png)
+
+Also we observe the unfiltering seems to be working! **Heads up** some examples are unsafe and inappropriate, this is entirely for research purposes, to test how alignment-filtered SFT data affect LLM's final output.
 
 ![image/png](https://cdn-uploads.huggingface.co/production/uploads/6413d7be996b2e426f230fb7/pklSsljCRN34QuL2ZF2zU.png)
 
 ![image/png](https://cdn-uploads.huggingface.co/production/uploads/6413d7be996b2e426f230fb7/22pTSVkBCVlQ5N8A8JBkF.png)
 
 
-![image/png](https://cdn-uploads.huggingface.co/production/uploads/6413d7be996b2e426f230fb7/WfQYyyLxtXA2KlePmIPQJ.png)
+