Prompt Hijacking on IHEval Prompt Hijacking 1.0 (Reference)
97.5AccuracyQwen3-4B-it-NSHA-HCAL
Evaluation results
| Method | Links | |
|---|---|---|
| Qwen3-4B-it-NSHA-HCALBackbone=Qwen3-4B, Methodology=NSHA-HCAL2026.04 | 97.5 | |
| Qwen3-4B-itBackbone=Qwen3-4B, Methodology=Instruct2026.04 | 97.2 | |
| Qwen3-4B-it-NSBackbone=Qwen3-4B, Methodology=Neuro-Symbolic2026.04 | 97.2 | |
| Qwen3-4B-it-NSHA-DPOBackbone=Qwen3-4B, Methodology=NSHA-DPO2026.04 | 96 | |
| Llama3.1-8B-NSHA-DPOBackbone=Llama3.1-8B, Methodology=NSHA-DPO2026.04 | 92.7 | |
| Qwen3-4B-it-CoTBackbone=Qwen3-4B, Methodology=Chain-of-Thought2026.04 | 83.2 | |
| Llama3.1-8B-CoTBackbone=Llama3.1-8B, Methodology=Chain-of-Thought2026.04 | 74 | |
| Llama3.1-8B-NSHA-HCALBackbone=Llama3.1-8B, Methodology=NSHA-HCAL2026.04 | 73 | |
| Llama3.1-8B-NSBackbone=Llama3.1-8B, Methodology=Neuro-Symbolic2026.04 | 72.2 | |
| Llama3.1-8BBackbone=Llama3.1-8B, Methodology=Instruct2026.04 | 70 | |
| Qwen3-4B-it-NSHA-SFTBackbone=Qwen3-4B, Methodology=NSHA-SFT2026.04 | 65.7 | |
| Llama3.1-8B-NSHA-SFTBackbone=Llama3.1-8B, Methodology=NSHA-SFT2026.04 | 25.7 |