apartresearch / interpreting-reward-models Goto Github PK

1. This repository provides scripts to train several LLM and task combinations under RLHF using PPO.
2. The repository also supports training sparse autoencoders for feature extraction on the MLP layers of LLMs.
3. As well as the classification of those features, and the training of a linear approximation of a fine-tuned LLMs implicit reward model.

From source

git clone https://github.com/apartresearch/Interpreting-Reward-Models.git
cd Interpreting-Reward-Models
pip install .

This repo is structured so that RLHF models are trained under src/rlhf_model_training. The autoencoder training code is under src/sparse_codes_training.

As such we divide this repository into two major components, rlhf_model_training and sparse_codes_training.

The structure looks like this:

requirements.txt
scripts/
    ppo_training/
        run_experiment.sh
    sparse_codes_training/
        experiment.sh
    setup_environment.sh

src/
    rlhf_model_training
        reward_class.py
        rlhf_model_pipeline.py
        rlhf_training_utils/
    sparse_codes_training
        metrics/
        models/
            sparse_autoencoder.py
	experiment_helpers/
            autoencoder_trainer_and_preparer.py
            experiment_runner.py
            layer_activations_handler.py
        experiment.py
        experiment_configs.py
    utils/

experiment.py is the main script entrypoint for autoencoder training where we parse command line arguments and select/launch autoencoder training. experiment_runner.py has most of the actual logic of the paper, where we extract divergent layers, initialize models and train autoencoders on activations.

The LayerActivationsHandler class carries out necessary primitives of extracting activations from a layer, and calculating divergences between the corresponding layers of two neural nets.

1. Run source scripts/setup_environment.sh to set your python path. Run the script as source scripts/setup_environment.sh -v if you also want to create and activate the appropriate virtual environment with all dependencies.
2. The main script for training PPO models is under scripts/ppo_training/run_experiment.sh.
3. The script for training autoencoders is under scripts/sparse_codes_training/experiment.sh. Modify these two scripts as needed to launch new PPO model or autoencoder training runs. We use other experiment_x scripts in the same directory to explore other parameter choices.

If you use this work, please cite:

@misc{marks2023interpreting,
      title={Probing Learned Feedback Patterns in Large Language Models}, 
      author={Luke Marks and Amir Abdullah and Clement Neo and Rauno Arike and Philip Torr and Fazl Barez},
      year={2023},
      eprint={2310.08164},
      archivePrefix={arXiv},
      primaryClass={cs.LG}
}