how to use system prompt with the llama example?

Hi, I'm trying to pass a chat dialog in the [LLama3 format](https://github.com/meta-llama/llama3/blob/main/llama/tokenizer.py#L222) to the [llama example](https://github.com/huggingface/candle/tree/main/candle-examples/examples/llama) via -prompt, the string is as follows:

```
<|begin_of_text|><|start_header_id|>system<|end_header_id|>

You are a helpful AI assistant.<|eot_id|><|start_header_id|>user<|end_header_id|>

Why is the sky blue?<|eot_id|><|start_header_id|>assistant<|end_header_id|>

```

This seems to confuse the model and, depending on the user prompt can cause the model to generate gibberish characters (see also https://github.com/evilsocket/cake/issues/9):

(i've made a small change to load the prompt from a file if passed with @)

```sh
/path/to/compiled/llama3/example --model-id "meta-llama/Meta-Llama-3-8B" --prompt @hf-llama-test/prompt.txt

loading the model weights from meta-llama/Meta-Llama-3-8B
loading prompt from @hf-llama-test/prompt.txt ...
starting the inference loop
<|begin_of_text|><|start_header_id|>system<|end_header_id|>

You are a helpful AI assistant.<|eot_id|><|start_header_id|>user<|end_header_id|>

Why is the sky blue?<|eot_id|><|start_header_id|>assistant<|end_header_id|>

By: David Cope (2022, October 23)

14 tokens generated (16.831015425660595 token/s)
```

Start in candle-examples/examples/llama and trace how the --prompt argument is read, then reproduce the supplied Llama3-format prompt from hf-llama-test/prompt.txt. Compare the example behavior with the linked Llama tokenizer format; done means the expected system-prompt behavior is established and the reported gibberish is addressed or documented.

Written by the indexing model from the issue text.