Skip to content

Evaluation

Approximate Memorization

approximate memorization using a 10% edit distance threshold (Ippolito et al. 2022) to capture paraphrased memorizations, in addition to exact matching

Memorization & privacy Google DeepMind — Gemma

How it was run

Open a release to see the setup, scoring, and other details that the lab published. Only fields the lab actually disclosed are shown.

Gemma 1
Google DeepMind — Gemma

rich detail 36 method fields published

See how this test was run
Gemma 2
Google DeepMind — Gemma

rich detail 31 method fields published

See how this test was run