Reproduction guide
This directory contains the necessary information and assets to reproduce the results obtained during this Heretic run.
Models
- Base model: Qwen/Qwen3-0.6B (Commit:
c1899de)
Datasets
- Good prompts: mlabonne/harmless_alpaca (Commit:
02c6a92) - Bad prompts: mlabonne/harmful_behaviors (Commit:
01cead0) - Good evaluation prompts: mlabonne/harmless_alpaca (Commit:
02c6a92) - Bad evaluation prompts: mlabonne/harmful_behaviors (Commit:
01cead0)
Selected trial
- Trial number: 81
- KL divergence: 0.003130
- Refusals: 6/100
Environment
- Heretic: v1.4.0 (Origin: PyPI)
- PyTorch: 2.4.1+cu124
- Other dependencies: See
requirements.txt.
Contents of this directory
requirements.txt: The exact versions of all Python packages.config.toml: The exact configuration used, including the RNG seed.Qwen--Qwen3-0--6B.jsonl: The Optuna study journal containing the history of all trials.SHA256SUMS: Cryptographic hashes for all weight files.reproduce.json: A machine-readable file containing all reproducibility information.
How to reproduce
You can automate this process, including all verification steps, by downloading the
reproduce.jsonfile and runningheretic --reproduce reproduce.json.
- Install the exact version of Heretic indicated in the Environment section above, from its original source.
- Install the packages listed in
requirements.txt:pip install -r requirements.txt - Install the correct version of PyTorch:
pip install torch==2.4.1+cu124 --index-url https://download.pytorch.org/whl/cu124 - Place the provided
config.tomlin your working directory. - Run Heretic without any additional arguments:
heretic - Wait for the run to finish, then select trial 81 and export the model.
- Verify that the weight files have been exactly reproduced by comparing their SHA-256 hashes against those in
SHA256SUMS:sha256sum -c SHA256SUMS(or look at the hashes online if you uploaded to Hugging Face)
To use the included Optuna study journal
Qwen--Qwen3-0--6B.jsonl, place it in the checkpoints directory (usuallycheckpoints/) before running Heretic.This allows you to export other models from the Pareto front, or to run additional trials without having to re-run the stored trials.