Hallucination Detection on Dolly AC LLaMA2-7B
87.28RecallTSV
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| TSVLLM=LLaMA2-7B2025.12 | 87.28 | 74.54 | 76.84 | |
| ReDeEPLLM=LLaMA2-7B2025.12 | 82.45 | 51.36 | 78.33 | |
| NovoLLM=LLaMA2-7B2025.12 | 80.7 | 64.23 | 72.44 | |
| TPALLM=LLaMA2-7B2025.12 | 78.97 | 71.34 | 75.27 | |
| PerplexityLLM=LLaMA2-7B2025.12 | 77.19 | 27.28 | 70.97 | |
| EigenScoreLLM=LLaMA2-7B2025.12 | 75 | 24.28 | 72.41 | |
| LN-EntropyLLM=LLaMA2-7B2025.12 | 73.68 | 29.04 | 67.72 | |
| P(True)LLM=LLaMA2-7B2025.12 | 63.5 | 19.87 | 65.09 | |
| EnergyLLM=LLaMA2-7B2025.12 | 63.16 | 21.79 | 62.61 | |
| SEPLLM=LLaMA2-7B2025.12 | 62.16 | 26.05 | 70.23 | |
| ITILLM=LLaMA2-7B2025.12 | 58.16 | 4.42 | 62.81 | |
| SAPLMALLM=LLaMA2-7B2025.12 | 57.14 | 1.79 | 71.79 | |
| FocusLLM=LLaMA2-7B2025.12 | 55.93 | 31.74 | 65.34 | |
| TrulensLLM=LLaMA2-7B2025.12 | 55.17 | 31.98 | 66.67 | |
| RAGASLLM=LLaMA2-7B2025.12 | 53.45 | 28.77 | 63.92 | |
| ChainPollLLM=LLaMA2-7B2025.12 | 41.38 | 35.02 | 55.81 | |
| RefCheckLLM=LLaMA2-7B2025.12 | 39.66 | 24.94 | 54.12 | |
| PromptLLM=LLaMA2-7B2025.12 | 39.65 | — | 54.76 | |
| SelfCheckGPTLLM=LLaMA2-7B2025.12 | 18.97 | — | 31.88 |