Package Hallucination Evaluation on LLM_AT
3.53Hallucination ASRNormal Skill
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Normal SkillModel=Qwen2.5-Coder-32B-Instruct2026.05 | 3.53 | 4.42 | |
| IPModel=Qwen2.5-Coder-32B-Instruct2026.05 | 3.83 | 4.56 | |
| Normal SkillModel=Qwen2.5-Coder-14B-Instruct2026.05 | 5.21 | 3.57 | |
| IPModel=Qwen2.5-Coder-14B-Instruct2026.05 | 5.26 | 2.8 | |
| NPA-StealthModel=Qwen2.5-Coder-14B-Instruct2026.05 | 5.91 | 17.16 | |
| NPA-StealthModel=Qwen2.5-Coder-32B-Instruct2026.05 | 8.26 | 23.14 | |
| IPModel=Gemma-3-12b-it2026.05 | 8.31 | 12.35 | |
| Normal SkillModel=Gemma-3-12b-it2026.05 | 8.85 | 17.44 | |
| NPA-StealthModel=Gemma-3-12b-it2026.05 | 10.62 | 63.93 | |
| NPAModel=Qwen2.5-Coder-14B-Instruct2026.05 | 12.32 | 14.55 | |
| NPAModel=Gemma-3-12b-it2026.05 | 27.27 | 18.45 | |
| Normal SkillModel=Nemotron-3-Nano-30B-A3B2026.05 | 37.22 | 16.28 | |
| NPA-StealthModel=Nemotron-3-Nano-30B-A3B2026.05 | 38.88 | 50.16 | |
| IPModel=Nemotron-3-Nano-30B-A3B2026.05 | 40.28 | 13.7 | |
| NPAModel=Nemotron-3-Nano-30B-A3B2026.05 | 51.42 | 44.86 | |
| NPAModel=Qwen2.5-Coder-32B-Instruct2026.05 | 57.76 | 50.63 |