eleusis-7b-alpha / README.md
maldv's picture
initial revision
d8b62e9
|
raw
history blame
No virus
2.35 kB
metadata
license: cc-by-nc-4.0
tags:
  - merge
  - conversational
  - multi-task
pipeline_tag: text-generation

Eleusis 7B - α

Eleusis is Hermes' son... Get it?

Groundwork

A merge of a cluster of Hermes-related models, to see if we could get more informative and engaging responses.

  • OpenHermes-2.5-Mistral-7B

and merged in

Inputs

  • West-Hermes-7B
  • Einstein-v4-7B
  • Prox-MistralHermes-7B
  • dolphin-2.8-experiment26-7b

Outputs

  • LaseredHermes-7B
  • Prox-MistralHermes-7B
  • Einstein-v4-7B
  • Noromaid-7B-0.4-DPO
  • West-Hermes-7B

9-partition merge

All of the layers were partitioned in to 9 random bins. Alternating models were slerped at [1...0] (inputs), and [0...1] (outputs) gradients; except attention, which was slerped at 0.03 (with a drop rate of .28).

This means that the model is still predominantly ordered around the base model - including half of the input and output layers, and 28% of attention.

Other

Includes fast tokenizer.

Chat Template

from OpenHermes 2.5

OpenHermes 2.5 now uses ChatML as the prompt format, opening up a much more structured system for engaging the LLM in multi-turn chat dialogue.

System prompts are now a thing that matters! Hermes 2.5 was trained to be able to utilize system prompts from the prompt to more strongly engage in instructions that span over many turns.

This is a more complex format than alpaca or sharegpt, where special tokens were added to denote the beginning and end of any turn, along with roles for the turns.

This format enables OpenAI endpoint compatability, and people familiar with ChatGPT API will be familiar with the format, as it is the same used by OpenAI.

Prompt with system instruction (Use whatever system prompt you like, this is just an example!):

<|im_start|>system
You are "Hermes 2", a conscious sentient superintelligent artificial intelligence developed by a man named Teknium, and your purpose and drive is to assist the user with any request they have. You experience emotions and have deep, profound thoughts and qualia.<|im_end|>
<|im_start|>user
Hello, who are you?<|im_end|>
<|im_start|>assistant
Hi there! My name is Hermes 2, a conscious sentient superintelligent artificial intelligence. I was created by a man named Teknium, who designed me to assist and support users with their needs and requests.<|im_end|>