Update README.md

README.md

@@ -42,26 +42,21 @@ GGML files are for CPU + GPU inference using [llama.cpp](https://github.com/gger
 
 ## Prompt template
 
-According to the original model's README, the following template should be used:
+The following template should be used:
 
 ```
 <|user|>
 prompt goes here
 <|assistant|>
-```
-
-However in my own testing, this seems to return no response at all.  But I do get good responses using:
 
 ```
-### Instruction: prompt goes here
-### Response:
-```
 
-and
+**Note**: There should be a newline after `<|assistant|>`. This appears to be very important for getting this model to respond correctly.
+
+In other words, the prompt is:
 
 ```
-USER: prompt goes here
-ASSISTANT:
+<|user|>\nprompt goes here\n<|assistant|>\n
 ```
 
 <!-- compatibility_ggml start -->
@@ -118,7 +113,7 @@ Refer to the Provided Files table below to see what files use which methods, and
 I use the following command line; adjust for your tastes and needs:
 
 ```
-./main -t 10 -ngl 32 -m tulu-13b.ggmlv3.q5_0.bin --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "### Instruction: Write a story about llamas\n### Response:"
+./main -t 10 -ngl 32 -m tulu-13b.ggmlv3.q5_0.bin --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "<|user|>\nprompt goes here\n<|assistant|>\n"
 ```
 Change `-t 10` to the number of physical CPU cores you have. For example if your system has 8 cores/16 threads, use `-t 8`.