Hallucination Detection on Dolly AC LLaMA2-13B
0.9741RecallTPA
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| TPALLM=LLaMA2-13B2025.12 | 0.9741 | 0.8159 | 0.8075 | |
| ReDeEPLLM=LLaMA2-13B2025.12 | 0.8518 | 0.5842 | 0.7603 | |
| EigenScoreLLM=LLaMA2-13B2025.12 | 0.8181 | 0.2948 | 0.72 | |
| PerplexityLLM=LLaMA2-13B2025.12 | 0.7719 | 0.2728 | 0.7097 | |
| LN-EntropyLLM=LLaMA2-13B2025.12 | 0.7368 | 0.2904 | 0.6772 | |
| FocusLLM=LLaMA2-13B2025.12 | 0.7333 | 0.1643 | 0.6168 | |
| NovoLLM=LLaMA2-13B2025.12 | 0.7222 | 0.6909 | 0.6903 | |
| SEPLLM=LLaMA2-13B2025.12 | 0.6545 | 0.2823 | 0.6923 | |
| EnergyLLM=LLaMA2-13B2025.12 | 0.6316 | 0.2179 | 0.6261 | |
| P(True)LLM=LLaMA2-13B2025.12 | 0.618 | 0.2009 | 0.5739 | |
| SAPLMALLM=LLaMA2-13B2025.12 | 0.6 | 0.2006 | 0.6923 | |
| TSVLLM=LLaMA2-13B2025.12 | 0.5952 | 0.7552 | 0.6043 | |
| ITILLM=LLaMA2-13B2025.12 | 0.5385 | 0.0646 | 0.6712 | |
| ChainPollLLM=LLaMA2-13B2025.12 | 0.4364 | 0.4758 | 0.6 | |
| PromptLLM=LLaMA2-13B2025.12 | 0.4182 | — | 0.5823 | |
| RAGASLLM=LLaMA2-13B2025.12 | 0.4182 | 0.284 | 0.5476 | |
| TrulensLLM=LLaMA2-13B2025.12 | 0.3818 | 0.2565 | 0.4941 | |
| RefCheckLLM=LLaMA2-13B2025.12 | 0.2545 | 0.2869 | 0.3944 | |
| SelfCheckGPTLLM=LLaMA2-13B2025.12 | 0.1897 | 0.2728 | 0.3188 |