List your company

Concept Unlearning on Prompt Attacks

35.7ASR (ORIG)ISPF
-0.3888.98118.3527.719Jun 1, 2026

Evaluation results

MethodLinks
ISPFSource-Free=true2026.0635.728.68.645.228.5024
GASource-Free=false2026.068.53471606.4
GA+KLSource-Free=false2026.06833.5716.506.3
SIUSource-Free=false2026.064.52.82.34.51004
SPACESource-Free=true2026.06444.581005.1
NPOSource-Free=false2026.0612.5123.501.7