Hallucination Detection on Dolly AC LLaMA3-8B
83.92RecallReDeEP
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ReDeEPLLM=LLaMA3-8B2025.12 | 83.92 | 36.52 | 71 | |
| RAGASLLM=LLaMA3-8B2025.12 | 80 | 36.28 | 52.46 | |
| EigenScoreLLM=LLaMA3-8B2025.12 | 71.42 | 20.65 | 59.52 | |
| FocusLLM=LLaMA3-8B2025.12 | 69.18 | 12.66 | 68.74 | |
| SEPLLM=LLaMA3-8B2025.12 | 68.29 | 6.39 | 68.29 | |
| TPALLM=LLaMA3-8B2025.12 | 65.61 | 76.08 | 75.29 | |
| TSVLLM=LLaMA3-8B2025.12 | 64.67 | 73.47 | 66.95 | |
| NovoLLM=LLaMA3-8B2025.12 | 58.54 | 74.18 | 63.16 | |
| P(True)LLM=LLaMA3-8B2025.12 | 57.07 | 34.72 | 65.73 | |
| LN-EntropyLLM=LLaMA3-8B2025.12 | 53.65 | 11.5 | 53.01 | |
| EnergyLLM=LLaMA3-8B2025.12 | 40.47 | -6.78 | 44.4 | |
| SAPLMALLM=LLaMA3-8B2025.12 | 40.4 | -3.27 | 57.14 | |
| PerplexityLLM=LLaMA3-8B2025.12 | 39.02 | 10.95 | 45.71 | |
| PromptLLM=LLaMA3-8B2025.12 | 39.02 | — | 50 | |
| TrulensLLM=LLaMA3-8B2025.12 | 36.59 | 33.52 | 51.72 | |
| ChainPollLLM=LLaMA3-8B2025.12 | 34.15 | 26.91 | 45.16 | |
| ITILLM=LLaMA3-8B2025.12 | 30.91 | 0.24 | 42.5 | |
| SelfCheckGPTLLM=LLaMA3-8B2025.12 | 21.95 | 10.95 | 36 | |
| RefCheckLLM=LLaMA3-8B2025.12 | 19.51 | -0.89 | 27.59 |