Llama3-Prime

Llama3-Prime

This Llama 3 8B model is a merge of other pretrained Llama 3 language models that were optimized for user preference. As a result, this merged model should be strong at providing relevant answers to user queries. Here, usability is more important than beating benchmarks.

• Input: text only
• Output: text only
• Prompt format: Llama 3
• Language: English

This model was created by merging multiple models with equal weights through the use of MergeKit's model_stock method.

Base Model: Daredevil-8B

Models Used:

• Llama-3-Instruct-8B-SimPO-ExPO
• Llama-3-8B-Magpie-Pro-SFT-v0.1
• SELM-Llama-3-8B-Instruct-iter-3
• LLaMA3-iterative-DPO-final-ExPO
• Llama-3-Instruct-8B-SPPO-Iter3
• MAmmoTH2-8B-Plus
• Bagel-8b-v1.0

Training Details:

The merged model was trained using LLaMA Factory on the alpaca_en_demo dataset to ensure the model can respond in the Llama 3 Instruct format. The training parameters included a rank of 1, an alpha value of 1, and a 0.3 dropout rate. In other words, very weak training to prevent interfering with the merged model's capabilities.