ao9000/llama2-7b-cda-race

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": 2.566261678495529,
    "total_flos": 2.605973017460736e+18,
    "train_loss": 0.7921548962593079,
    "train_runtime": 150925.8962,
    "train_samples": 24939,
    "train_samples_per_second": 0.424,
    "train_steps_per_second": 0.013
}

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