Hallucination Detection on RAGTruth LLaMA2-13B
80.68RecallTSV
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| TSVLLM=LLaMA2-13B2025.12 | 80.68 | 81.23 | 69.87 | |
| TPALLM=LLaMA2-13B2025.12 | 79.13 | 86.81 | 79.75 | |
| ChainPollLLM=LLaMA2-13B2025.12 | 78.74 | 74.14 | 73.42 | |
| NovoLLM=LLaMA2-13B2025.12 | 78.26 | 85.06 | 77.33 | |
| TrulensLLM=LLaMA2-13B2025.12 | 77.29 | 70.73 | 68.67 | |
| ReDeEPLLM=LLaMA2-13B2025.12 | 71.98 | 82.44 | 75.87 | |
| PromptLLM=LLaMA2-13B2025.12 | 70 | — | 68.99 | |
| RefCheckLLM=LLaMA2-13B2025.12 | 68 | 78.57 | 70.23 | |
| RAGASLLM=LLaMA2-13B2025.12 | 67.63 | 75.41 | 67.47 | |
| EigenScoreLLM=LLaMA2-13B2025.12 | 67.15 | 66.4 | 66.37 | |
| SEPLLM=LLaMA2-13B2025.12 | 65.8 | 80.89 | 71.59 | |
| FocusLLM=LLaMA2-13B2025.12 | 61.73 | 78.88 | 69.77 | |
| P(True)LLM=LLaMA2-13B2025.12 | 59.8 | 79.98 | 70.32 | |
| ITILLM=LLaMA2-13B2025.12 | 55.19 | 80.51 | 68.38 | |
| LN-EntropyLLM=LLaMA2-13B2025.12 | 53.83 | 59.12 | 66.55 | |
| PerplexityLLM=LLaMA2-13B2025.12 | 51.9 | 50.91 | 67.49 | |
| EnergyLLM=LLaMA2-13B2025.12 | 50.57 | 56.19 | 66.57 | |
| SAPLMALLM=LLaMA2-13B2025.12 | 50.53 | 80.29 | 65.29 | |
| SelfCheckGPTLLM=LLaMA2-13B2025.12 | 46.42 | — | 46.42 |