This is a my first Llama 3 MoE model with the following configs:

base_model: Llama-3-RPMerge-8B-SLERP

experts:

  - source_model: Llama-3-RPMerge-8B-SLERP
  - source_model: WesPro_Daring_Llama
  - source_model: Chaos_RP_l3_8B
  - source_model: llama-3-stinky-8B

It's meant for RP and does pretty well at it but I haven't tested it excessively yet.