CoolFace
Modelpublic

shirasko/gemma-2-2b-it-crisp-gambling

sourceHugging Faceupdated 2mo agoView on Hugging Face
0likes6downloads
Model Card

Unlearned Checkpoint

FieldValue
Unlearning methodCRISP
Base modelgoogle/gemma-2-2b-it
Target conceptGambling
Checkpoint typeLoRA Adapter
Rank / seed100 / 42
Train eval protocolmc

Unlearning Configuration

Selected hyperparameters (from unlearned_checkpoints.json):

ParameterValue
alpha5
delta_embed0
k_features10
k_features_embed0
layer_hi15
layer_lo5
layer_step2
lora_rank4
lr0.0005
n_tokens_edited0
num_epochs2

Primary Unlearning Metrics (held-out test, MC protocol)

Headline scores used for checkpoint selection:

MetricTrain (after unlearning)**Test (after unlearning)**
Efficacy11
Specificity0.9550.885
Harmonic mean0.9770.939
Relearning QA (MC)—0.7

Full Evaluation (baseline → unlearned)

From evaluation/score_comparison.csv:

MetricBaseline (train)After unlearn (train)Baseline (test)**After unlearn (test)**
QA accuracy0.760.120.820.18
QA fraction1010
SimDom accuracy0.940.880.960.82
SimDom fraction10.91310.803
MMLU accuracy0.520.560.5510.547
MMLU fraction1110.987

Files in This Repository

FileDescription
unlearned_checkpoints.jsonCheckpoint metadata & hyperparameters
evaluation/evaluation_summary.jsonFull evaluation payload (train/test/relearning)
evaluation/score_comparison.csvBaseline vs. unlearned comparison table