hugpadhugpad

ContextualAI / archangel_sft-ppo_pythia2-8b

Text Generation2.8B310Updated Jan 11, 2024
Hugging Face

Downloads

31

Likes

0

Last updated

Jan 11, 2024

License

apache-2.0

This repo contains the model checkpoints for: - model family <bEleutherAI/pythia-2.8b</b - optimized with the loss <bPPO</b - aligned using the SHP, Anthropic HH and Open Assistant datasets.

Tasks & library

  • text generation
  • transformers

Tags

  • transformers
  • gpt_neox
  • text-generation
  • human feedback
  • rlhf
  • preferences

Technical specifications

Parameters2.8B
Librarytransformers
Languagesen
Licenseapache-2.0