ao9000/llama2-7b-cda-gender

Lora finetuning on Wikipedia-10, applying counter factual data augmentation (CDA)

• Dataset: Wikipedia-10
• Target modules = ["q_proj", "k_proj", "v_proj", "o_proj", "gate_proj", "up_proj", "down_proj"]

{
    "epoch": 0.05217151712326255,
    "total_flos": 2.605973017460736e+18,
    "train_loss": 0.03771647383272648,
    "train_runtime": 18281.2091,
    "train_samples": 1226723,
    "train_samples_per_second": 3.501,
    "train_steps_per_second": 0.109
}

Training script: https://github.com/ao9000/bias-bench/blob/main/experiments/run_clm.py