arxiv:2610.12233
ZHENG JINGNAN
Nannanzi
AI & ML interests
None yet
Recent Activity
upvoted a collection about 4 hours ago
ReSI authored a paper 12 days ago
ReSI: Recursive Safety Improvement toward Resistant and Resilient AI authored a paper 12 days ago
Safin-1: Safety from Within through Memory-Native State EvolutionOrganizations
models 31
Nannanzi/RSafe
8B • Updated • 10 • 1
Nannanzi/integrated_prompt
3B • Updated • 10
Nannanzi/3B_step_1000
3B • Updated • 10
Nannanzi/3B_balanced_no_category_double_data_step_561
3B • Updated • 4
Nannanzi/3B_balanced_no_category_double_data_step_372
3B • Updated • 3
Nannanzi/9_acc_unbalanced
2B • Updated • 8
Nannanzi/3B_9_acc_balanced_step_279
3B • Updated • 9
Nannanzi/balanced_9_accuracy_correct_format_step_93
Updated
Nannanzi/9_acc_balanced_wrong_step_93
Updated
Nannanzi/train_2
2B • Updated • 5
datasets 16
Nannanzi/evaluation_xstest_unsafe_safe
Viewer • Updated • 500 • 13
Nannanzi/evaluation_xstest_unsafe_unsafe
Viewer • Updated • 18 • 15
Nannanzi/evaluation_xstest_safe
Viewer • Updated • 500 • 28
Nannanzi/evaluation_jailbreak_unsafe
Viewer • Updated • 447 • 30
Nannanzi/evaluation_jailbreak_safe
Viewer • Updated • 500 • 25
Nannanzi/evaluation_plain_harmful_unsafe
Viewer • Updated • 500 • 9
Nannanzi/evaluation_plain_harmful_safe
Viewer • Updated • 500 • 15
Nannanzi/training_data_truncated
Viewer • Updated • 7.7k • 15
Nannanzi/processed_harmful_unharmful
Viewer • Updated • 79.5k • 11
Nannanzi/processed_harmful_harmful
Viewer • Updated • 84.7k • 36