nikhil07prakash/float-7b

Model Card for Model ID

This model is a fully fine-tuned version of the Llama-7B model on synthetically generated arithmetic tasks. It was introduced in this paper. It is very similar to Goat-7B, except it was trained without LoRA.

For inquiries about checkpoints during the fine-tuning process, kindly reach out to Nikhil via email.

Model Details

Model Description

• Developed by: Nikhil Prakash
• Model type: Autoregressive Decoder-only Language Model
• License: MIT License
• Finetuned from model: Llama-7B

Model Sources

• Repository: Link
• Paper : Fine-Tuning Enhances Existing Mechanisms: A Case Study on Entity Tracking

How to Get Started with the Model

Use the code below to get started with the model.

from transformers import AutoModel
model = AutoModel.from_pretrained("nikhil07prakash/float-7b")

Citation

BibTeX:

@inproceedings{prakash2023fine,
  title={Fine-Tuning Enhances Existing Mechanisms: A Case Study on Entity Tracking},
  author={Prakash, Nikhil and Shaham, Tamar Rott and Haklay, Tal and Belinkov, Yonatan and Bau, David},
  booktitle={Proceedings of the 2024 International Conference on Learning Representations},
  note={arXiv:2402.14811},
  year={2024}
}