Hermes 2 Theta Llama 3 70B
Hermes 2 Theta Llama 3 70B combines Hermes 2 Pro and Meta's Llama 3 Instruct, followed by reinforcement learning from human feedback.
Research release
Training and model behavior
Nous developed the experimental merge with Charles Goddard and Arcee AI. It uses ChatML for conversation and structured prompts. It is a distinct trained merge, rather than a file-format conversion.Hermes 2 Theta Llama 3 70B model card
Working with this model
Practical guidance. Compare the merge with both parent models using the same instructions. Mixing weights can change style and refusal behavior as well as task accuracy.
Download formats
Nous also publishes GGUF packages for this model. These are deployment formats of the same release.Hermes 2 Theta Llama 3 70B GGUF package
Specifications & access
| Detail | Value and conditions |
|---|---|
| Status | Research release |
| Launch date | |
| Access | Downloadable |
| Model identifiers | NousResearch/Hermes-2-Theta-Llama-3-70B |
| Context window | 8,192 tokens Configured position limit; practical serving limits depend on the runtime. Hermes 2 Theta Llama 3 70B configuration |
| Base model | Hermes 2 Pro and Meta Llama 3 InstructHermes 2 Theta Llama 3 70B model card |
| Repository license label | llama3Hermes 2 Theta Llama 3 70B model card |
Related models
- Hermes 2 Theta Llama 3 8B
Nous Research ยท Hermes 2 Theta
A merged model with additional preference training.
Launch date May 2024Research releaseDownloadable