nonetrix/sillyrp-7b

Silly RP 7B

This is one of my first merges, just throwing random things at the wall to see what works. Seems really good in my limited testing, but I can't make any promisses of it's quality. Just try it yourself and give me feedback I guess, I am rather intrested to see what others think about it. I don't know what I am doing, I don't even know benifits of diffrent merge methods :P

General setting guide (will update over time, share your settings please):

• Chat format ChatML (all the models merged are ChatML)
• Needs high repetition penalty after a while, starts out good but slowly gets into a loop without it

---

base_model:

• tavtav/eros-7b-test
• NousResearch/Nous-Hermes-2-Mistral-7B-DPO
• maywell/Synatra-7B-v0.3-RP
• NeverSleep/Noromaid-7B-0.4-DPO
• cogbuji/Mr-Grammatology-clinical-problems-Mistral-7B-0.5 library_name: transformers tags:
• mergekit
• merge

---

output

This is a merge of pre-trained language models created using mergekit.

Merge Details

Merge Method

This model was merged using the task arithmetic merge method using NeverSleep/Noromaid-7B-0.4-DPO as a base.

Models Merged

The following models were included in the merge:

• tavtav/eros-7b-test
• NousResearch/Nous-Hermes-2-Mistral-7B-DPO
• maywell/Synatra-7B-v0.3-RP
• cogbuji/Mr-Grammatology-clinical-problems-Mistral-7B-0.5

Configuration

The following YAML configuration was used to produce this model:

base_model: NeverSleep/Noromaid-7B-0.4-DPO
models:
      - model: maywell/Synatra-7B-v0.3-RP
        parameters:
          weight: 0.2
      - model: tavtav/eros-7b-test
        parameters:
          weight: 0.2
      - model: cogbuji/Mr-Grammatology-clinical-problems-Mistral-7B-0.5
        parameters:
          weight: 0.2
      - model: NousResearch/Nous-Hermes-2-Mistral-7B-DPO
        parameters:
          weight: 0.2
merge_method: task_arithmetic
parameters:
      weight: 0.17
dtype: float16
random_seed: 694201337567099116663322537