# Can a training input end with a user message?

**URL:** <https://discuss.huggingface.co/t/can-a-training-input-end-with-a-user-message/76608>\
**Category:** 🤗Datasets\
**Created:** [March 9, 2024, 1:34am UTC](https://discuss.huggingface.co/t/can-a-training-input-end-with-a-user-message/76608 "2024-03-09T01:34:47Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![capnchat](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/capnchat/32/31430_2.png) [@capnchat](https://discuss.huggingface.co/u/capnchat)\
**Post date:** [March 9, 2024, 1:34am UTC](https://discuss.huggingface.co/t/can-a-training-input-end-with-a-user-message/76608/1 "2024-03-09T01:34:47Z")

</div>

I am creating training inputs from data like this:

```auto
[
{ "role": "user", "content": "prompt 1" },
{ "role": "assistant", "content": "response 1" },
{ "role": "user", "content": "prompt 2" },
{ "role": "assistant", "content": "response 2" }
]

```

which will be converted to something like this for mistral’s instruct syntax:

```auto
<s>[INST] prompt 1 [/INST] response 1</s> [INST] prompt 2 [/INST] response 2</s>

```

My question is, can Mistral-7B-Instruct-v0.2 still learn from data such as the below, which ends with a user message instead of an assistant?

```auto
[
{ "role": "user", "content": "prompt 1" },
{ "role": "assistant", "content": "response 1" },
{ "role": "user", "content": "prompt 2" }
]

```

or…

```auto
<s>[INST] prompt 1 [/INST] response 1</s> [INST] prompt 2 [/INST]

```

**Could it still be useful to include the extra user message so the AI could learn from the token sequences in the final user prompt? Or is it a bad idea to include an instruction without a corresponding response?**

Thanks a lot for any insight you have into this.