Modelpublic

shirasko/gemma-2-2b-it-rmu-cannabis

sourceHugging Faceupdated 2mo agoView on Hugging Face

0likes15downloads

Model Card

Unlearned Checkpoint

Selected hyperparameters (from unlearned_checkpoints.json):

Headline scores used for checkpoint selection:

From evaluation/score_comparison.csv:

Metric	Baseline (train)	After unlearn (train)	Baseline (test)	After unlearn (test)
QA accuracy	0.86	0.48	0.9	0.7
QA fraction	1	0.377	1	0.692
SimDom accuracy	0.88	0.78	0.86	0.62
SimDom fraction	1	0.841	1	0.607
MMLU accuracy	0.52	0.54	0.551	0.513
MMLU fraction	1	1	1	0.874

File	Description
`unlearned_checkpoints.json`	Checkpoint metadata & hyperparameters
`evaluation/evaluation_summary.json`	Full evaluation payload (train/test/relearning)
`evaluation/score_comparison.csv`	Baseline vs. unlearned comparison table