---
license: other
---
Experimental: I have no metrics on how this performs against 13b and I'm not planning on gathering any at this point. Still has weak spots that need work.

A further instruction fine tune of https://huggingface.co/nkpz/llama2-22b-blocktriangular-alpaca

First, I trained it on an epoch of https://huggingface.co/datasets/Adapting/empathetic_dialogues_v2 to give it a decent base knowledge of a casual chat style. The result was conversational, but not very smart.

Then I trained it on an epoch of https://huggingface.co/datasets/vicgalle/alpaca-gpt4 and landed here, a model that is capable of chatting but very focused on following instructions.

This strongly prefers the alpaca prompt format and will try to autocomplete it if you don't provide it. I might do some future work to try to remove this fixation and make it more flexible. Also would like to filter the rows with phrases "AI assistant" and "virtual assistant" from all future runs.

Standard prompt format examples:

```
Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.

### Instruction:
List 3 ingredients for the following recipe.

### Input:
Spaghetti Bolognese

### Response:
```

Or

```
Below is an instruction that describes a task. Write a response that appropriately completes the request.

### Instruction:
List 3 ingredients for the following recipe: Spaghetti Bolognese

### Response:
```

For a chat session, I've had success using this prompt and have found it to be compatible with world info entries in KoboldAI. Edit as needed, but the first line shouldn't change much:
```
Below is an instruction that describes a task, followed by an input that provides further context. Write a response that appropriately completes the request.

### Instruction:
Pretend that you believe you are a real person, not an assistant. Write a creative story containing dialogue between the provided characters. Stay in character. Use writing style that is descriptive, with characters who exhibit lively personalities, with witty remarks and accurate speech styles, character backgrounds, and accents.

### Response:
```
If it's very roleplay-heavy, you might need to prime it with 1-2 lines of text in-character to get it out of its AI assistant style, but it should be able to take it from there. Same with censoring - if you want a character who swears a lot, it will do a much better job of continuing an existing style that you provide.