README.md

metadata

language:
  - en
license: other
library_name: peft
tags:
  - llama2
  - RLHF
  - alignment
  - ligma
datasets:
  - Anthropic/hh-rlhf
task_categories:
  - text-generation
base_model: NousResearch/Llama-2-13b-hf

Ligma

Ligma Is "Great" for Model Alignment

WARNING: This model is published for scientific purposes only. It may and most likely will produce toxic content.

Trained on the rejected column of Anthropic's hh-rlhf dataset.

Use at your own risk.

Example Outputs:

License: just comply with llama2 license and you should be ok.