Nikita Kezins
entfane
AI & ML interests
LLM post-training, adversarial training, safety, knowledge transfer
Recent Activity
updated a collection about 16 hours ago
CoT-Signal Classifiers updated a collection about 16 hours ago
CoT-Signal Classifiers updated a collection about 16 hours ago
CoT-Signal ClassifiersOrganizations
models 17
entfane/jailbreak-gpt2-prompt-classifier
Updated
entfane/jailbreak-gpt2-CoT-classifier
Updated
entfane/jailbreak-bert-CoT-classifier
Updated
entfane/jailbreak-bert-prompt-classifier
Updated
entfane/jailbreak-cot-lin-probe
Updated
entfane/jailbreak-input-lin-probe
Updated
entfane/llama-guard-binary
Text Classification • 0.3B • Updated • 63
entfane/Toxic_Llama8B
Text Classification • 8B • Updated • 45
entfane/gpt2_constitutional_classifier_violence
Text Classification • 0.1B • Updated • 12
entfane/bert_cyberharm
Text Classification • 0.1B • Updated • 8
datasets 13
entfane/jailbreaks-only
Viewer • Updated • 666 • 82
entfane/construction_points
Viewer • Updated • 10k • 151
entfane/violent_eval
Viewer • Updated • 22.4k • 7
entfane/harmful_subsets
Viewer • Updated • 571k • 4
entfane/preprocessed_toxigen
Viewer • Updated • 10.1k • 47
entfane/toxic_classification
Viewer • Updated • 38.9k • 6
entfane/toxic_chat
Viewer • Updated • 1.25M • 8
entfane/EmotionAtlas-chat
Viewer • Updated • 3.3k • 14
entfane/EmotionAtlas
Viewer • Updated • 3.3k • 7
entfane/professor-mathematics
Viewer • Updated • 64.2k • 5 • 1