Skip to content

Llama-3-Giraffe-70B-Instruct

A 70B model with a 128k effective context

Visit official site
Category AI models

What Llama-3-Giraffe-70B-Instruct is

The model is described as an Abacus.AI large language model trained with PoSE and dynamic NTK interpolation to reach a longer effective context.

What you can do with it

  • Generate long stretches of text
  • Work through large volumes of text data
  • Apply adapter transfer to improve performance

Getting started

  • Find the model on the Hugging Face platform
  • Set values such as context length and sample count
  • Upload or enter the text data
  • Run the model and inspect the output

Pros & cons

✓ What we like

  • Long text generation is supported
  • Evaluation uses a Needle-in-a-Haystack test

! What to watch out for

  • Parameters such as context length and sample count are set by the user
  • Text data must be uploaded or entered before running

FAQ

What context length is stated?

The card states about 128k effective context length.

Which methods are named?

The features name PoSE, dynamic NTK interpolation and adapter transfer.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review