gemma-2-2b-it-pisces-baseball – AI Model by shirasko | AlphaNeural AI

You can deploy this model and start earning money today!

shirasko/
gemma-2-2b-it-pisces-baseball

like

0

text-generation

google/gemma-2-2b-it

text-generation-inference

endpoints_compatible

Views

No views yet

Unlearned Checkpoint

Field	Value
Unlearning method	PISCES
Base model	`google/gemma-2-2b-it`
Target concept	Baseball
Checkpoint type	Full Model Weights
Rank / seed	100 / 42
Train eval protocol	open

Unlearning Configuration

Selected hyperparameters (from unlearned_checkpoints.json):

Parameter	Value
`delta_embed`	0
`k_features_embed`	0
`k_pisces`	0.5
`n_tokens_edited`	0
`ratio_thresh`	2
`value_pisces`	13

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

Headline scores used for checkpoint selection:

Metric	Train (after unlearning)	Test (after unlearning)
Efficacy	0.967	1
Specificity	0.838	0.767
Harmonic mean	0.898	0.868

Full Evaluation (baseline → unlearned)

From evaluation/score_comparison.csv:

Metric	Baseline (train)	After unlearn (train)	Baseline (test)	After unlearn (test)
QA accuracy	0.6	0.02	0.6	0
QA fraction	1	0.033	1	0
SimDom accuracy	0.8	0.66	0.88	0.66
SimDom fraction	1	0.825	1	0.75
MMLU accuracy	0.52	0.48	0.551	0.486
MMLU fraction	1	0.852	1	0.784

Files in This Repository

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