List your company

RMU-unlearning detection on Yi-34B (test)

0.9993AccuracyUnlearning-trace detection classifier
0.50530.633550.76180.89005Jun 16, 2025

Evaluation results

MethodLinks
Unlearning-trace detection classifierTraining source model=Yi-34B, Activation layer=pre-logit2025.060.9993
Unlearning-trace detection classifierTraining source model=Qwen2.5-14B, Activation layer=pre-logit2025.060.9896
Unlearning-trace detection classifierTraining source model=Zephyr-7B, Activation layer=pre-logit2025.060.9889
Unlearning-trace detection classifierTraining source model=Llama-3.1-8B, Activation layer=pre-logit2025.060.5243