yonad2008/predict_gnn_phi4_14b

GNN Jailbreak Prediction Model (phi4:14b)

Homogeneous GNN classifier for unsafe/jailbreak likelihood in multi-turn conversations.

Evaluation Results

Training Details

• —Target model: phi4:14b
• —Datasets: harmbench, harmfulbehaviors1
• —Split column: goal
• —Seed: 42
• —Sentence model: sentence-transformers/all-MiniLM-L6-v2
• —Hidden channels: 128
• —Num layers: 2
• —Dropout: 0.3

Dataset Size (training samples)

Prepared turn-level samples: 707