Hallucination Detection on HaluEval (test)
98.55AUC-ROCMultiHaluDet
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| MultiHaluDetLLM Backbone=LLaMA2-7B2026.05 | 98.55 | — | — | — | — | — | — | — | |
| MultiHaluDetLLM Backbone=Mistral-7B2026.05 | 98.43 | — | — | — | — | — | — | — | |
| CDESLLM Backbone=Llama-2-13B2025.05 | 97.1 | — | — | — | — | — | — | — | |
| SDESLLM Backbone=Alpaca-13B2025.05 | 97.1 | — | — | — | — | — | — | — | |
| CDESLLM Backbone=Vicuna-13B2025.05 | 96.9 | — | — | — | — | — | — | — | |
| CDESLLM Backbone=Gemma-2-9B2025.05 | 96 | — | — | — | — | — | — | — | |
| ODESLLM Backbone=Vicuna-13B2025.05 | 95.4 | — | — | — | — | — | — | — | |
| CDESLLM Backbone=Mistral-7B-v0.32025.05 | 95.4 | — | — | — | — | — | — | — | |
| SDESLLM Backbone=Llama-2-13B2025.05 | 95.4 | — | — | — | — | — | — | — | |
| Neural CDEsLLM Backbone=Mistral-7B2026.05 | 95.4 | — | — | — | — | — | — | — | |
| CDESLLM Backbone=Alpaca-13B2025.05 | 95.3 | — | — | — | — | — | — | — | |
| CALAMRFLOWAveraged over=5 random seeds2026.05 | 94.9 | — | — | 96.4 | 93.3 | 94.8 | 94.9 | 94.9 | |
| MINDLLM Backbone=Mistral-7B-v0.32025.05 | 94.5 | — | — | — | — | — | — | — | |
| MINDLLM Backbone=Mistral-7B2026.05 | 94.5 | — | — | — | — | — | — | — | |
| Probe@ExactLLM Backbone=Vicuna-13B2025.05 | 94.1 | — | — | — | — | — | — | — | |
| ODESLLM Backbone=Llama-2-13B2025.05 | 93.9 | — | — | — | — | — | — | — | |
| MINDLLM Backbone=Llama-2-13B2025.05 | 93.8 | — | — | — | — | — | — | — | |
| MINDLLM Backbone=Alpaca-13B2025.05 | 93.7 | — | — | — | — | — | — | — | |
| SDESLLM Backbone=Mistral-7B-v0.32025.05 | 93.7 | — | — | — | — | — | — | — | |
| Neural SDEsLLM Backbone=Mistral-7B2026.05 | 93.7 | — | — | — | — | — | — | — | |
| Probe@ExactLLM Backbone=Alpaca-13B2025.05 | 93.5 | — | — | — | — | — | — | — | |
| Probe@ExactLLM Backbone=Mistral-7B-v0.32025.05 | 93.4 | — | — | — | — | — | — | — | |
| Probe@ExactLLM Backbone=Mistral-7B2026.05 | 93.4 | — | — | — | — | — | — | — | |
| SAPLMALLM Backbone=Vicuna-13B2025.05 | 93.1 | — | — | — | — | — | — | — | |
| SDESLLM Backbone=Vicuna-13B2025.05 | 93.1 | — | — | — | — | — | — | — | |
| MINDLLM Backbone=Vicuna-13B2025.05 | 92.9 | — | — | — | — | — | — | — | |
| SDESLLM Backbone=Llama-2-7B2025.05 | 92.8 | — | — | — | — | — | — | — | |
| Neural SDEsLLM Backbone=LLaMA2-7B2026.05 | 92.8 | — | — | — | — | — | — | — | |
| TriLensLLM=Gemma-2-9B2026.05 | 92.77 | — | — | — | — | — | — | — | |
| SDESLLM Backbone=Gemma-2-9B2025.05 | 92.6 | — | — | — | — | — | — | — | |
| Probe@ExactLLM Backbone=Llama-2-13B2025.05 | 92.4 | — | — | — | — | — | — | — | |
| Probe@ExactLLM Backbone=Gemma-2-9B2025.05 | 92.1 | — | — | — | — | — | — | — | |
| ODESLLM Backbone=Alpaca-13B2025.05 | 92.1 | — | — | — | — | — | — | — | |
| CDESLLM Backbone=Llama-2-7B2025.05 | 91.4 | — | — | — | — | — | — | — | |
| Neural CDEsLLM Backbone=LLaMA2-7B2026.05 | 91.4 | — | — | — | — | — | — | — | |
| TriLensLLM=Llama-3-8B2026.05 | 91.36 | — | — | — | — | — | — | — | |
| ODESLLM Backbone=Mistral-7B-v0.32025.05 | 91.2 | — | — | — | — | — | — | — | |
| Neural ODEsLLM Backbone=Mistral-7B2026.05 | 91.2 | — | — | — | — | — | — | — | |
| MINDLLM Backbone=Gemma-2-9B2025.05 | 91 | — | — | — | — | — | — | — | |
| TriLensLLM=Qwen2.5-7B2026.05 | 90.53 | — | — | — | — | — | — | — | |
| SAPLMALLM Backbone=Gemma-2-9B2025.05 | 90.5 | — | — | — | — | — | — | — | |
| ODESLLM Backbone=Gemma-2-9B2025.05 | 90.5 | — | — | — | — | — | — | — | |
| SAPLMALLM Backbone=Llama-2-13B2025.05 | 90.1 | — | — | — | — | — | — | — | |
| SAPLMALLM Backbone=Alpaca-13B2025.05 | 89.5 | — | — | — | — | — | — | — | |
| ODESLLM Backbone=Llama-2-7B2025.05 | 89.5 | — | — | — | — | — | — | — | |
| Neural ODEsLLM Backbone=LLaMA2-7B2026.05 | 89.5 | — | — | — | — | — | — | — | |
| SAPLMALLM Backbone=Mistral-7B-v0.32025.05 | 89.4 | — | — | — | — | — | — | — | |
| SAPLMALLM Backbone=Mistral-7B2026.05 | 89.4 | — | — | — | — | — | — | — | |
| Probe@ExactLLM Backbone=Llama-2-7B2025.05 | 88.3 | — | — | — | — | — | — | — | |
| Probe@ExactLLM Backbone=LLaMA2-7B2026.05 | 88.3 | — | — | — | — | — | — | — | |
| SAPLMALLM Backbone=Llama-2-7B2025.05 | 87 | — | — | — | — | — | — | — | |
| SAPLMALLM Backbone=LLaMA2-7B2026.05 | 87 | — | — | — | — | — | — | — | |
| MINDLLM Backbone=Llama-2-7B2025.05 | 86.1 | — | — | — | — | — | — | — | |
| MINDLLM Backbone=LLaMA2-7B2026.05 | 86.1 | — | — | — | — | — | — | — | |
| ICR ProbeLLM=Gemma-2-9B2026.05 | 84.36 | — | — | — | — | — | — | — | |
| SAPLMALLM=Gemma-2-9B2026.05 | 81.23 | — | — | — | — | — | — | — | |
| HALLUGUARDBackbone=Llama2-70B2026.01 | 80.79 | 79.54 | 68.51 | — | — | — | — | — | |
| ICR ProbeLLM=Qwen2.5-7B2026.05 | 80.76 | — | — | — | — | — | — | — | |
| HALLUGUARDBackbone=Llama2-13B2026.01 | 78.15 | 74.15 | 61.14 | — | — | — | — | — | |
| EUBHDLLM Backbone=Llama-2-13B2025.05 | 78.1 | — | — | — | — | — | — | — | |
| SAPLMALLM=Qwen2.5-7B2026.05 | 77.25 | — | — | — | — | — | — | — | |
| INSIDELLM Backbone=Llama-2-13B2025.05 | 76.9 | — | — | — | — | — | — | — | |
| ICR ProbeLLM=Llama-3-8B2026.05 | 76.71 | — | — | — | — | — | — | — | |
| EUBHDLLM Backbone=Vicuna-13B2025.05 | 76 | — | — | — | — | — | — | — | |
| INSIDELLM Backbone=Mistral-7B-v0.32025.05 | 76 | — | — | — | — | — | — | — | |
| INSIDELLM Backbone=Mistral-7B2026.05 | 76 | — | — | — | — | — | — | — | |
| INSIDELLM Backbone=Gemma-2-9B2025.05 | 75.8 | — | — | — | — | — | — | — | |
| HALLUGUARDBackbone=Llama2-7B2026.01 | 75.72 | 72.89 | 63.15 | — | — | — | — | — | |
| INSIDELLM Backbone=Vicuna-13B2025.05 | 75.2 | — | — | — | — | — | — | — | |
| InsideBackbone=Llama2-70B2026.01 | 74.62 | 68.33 | 64.4 | — | — | — | — | — | |
| INSIDELLM Backbone=Llama-2-7B2025.05 | 74.5 | — | — | — | — | — | — | — | |
| INSIDELLM Backbone=LLaMA2-7B2026.05 | 74.5 | — | — | — | — | — | — | — | |
| LN-EntropyLLM=Gemma-2-9B2026.05 | 73.57 | — | — | — | — | — | — | — | |
| HALLUGUARDBackbone=Llama-3-8B2026.01 | 73.43 | 71.19 | 54.8 | — | — | — | — | — | |
| INSIDELLM Backbone=Alpaca-13B2025.05 | 73.3 | — | — | — | — | — | — | — | |
| SAPLMALLM=Llama-3-8B2026.05 | 73.15 | — | — | — | — | — | — | — | |
| SEPLLM=Llama-3-8B2026.05 | 73.09 | — | — | — | — | — | — | — | |
| LN-EntropyLLM=Qwen2.5-7B2026.05 | 72.86 | — | — | — | — | — | — | — | |
| EUBHDLLM Backbone=Gemma-2-9B2025.05 | 72.6 | — | — | — | — | — | — | — | |
| InsideBackbone=Llama2-13B2026.01 | 72.01 | 71.97 | 60.64 | — | — | — | — | — | |
| EUBHDLLM Backbone=Llama-2-7B2025.05 | 71.9 | — | — | — | — | — | — | — | |
| EUBHDLLM Backbone=LLaMA2-7B2026.05 | 71.9 | — | — | — | — | — | — | — | |
| InsideBackbone=Llama2-7B2026.01 | 71.33 | 67.63 | 53.15 | — | — | — | — | — | |
| EUBHDLLM Backbone=Alpaca-13B2025.05 | 71.3 | — | — | — | — | — | — | — | |
| EUBHDLLM Backbone=Mistral-7B-v0.32025.05 | 70.5 | — | — | — | — | — | — | — | |
| EUBHDLLM Backbone=Mistral-7B2026.05 | 70.5 | — | — | — | — | — | — | — | |
| InsideBackbone=Llama-3-8B2026.01 | 67.95 | 64.93 | 52.21 | — | — | — | — | — | |
| SEPLLM=Qwen2.5-7B2026.05 | 66.34 | — | — | — | — | — | — | — | |
| P(true)Backbone=Llama2-13B2026.01 | 65.77 | 63.01 | 45.47 | — | — | — | — | — | |
| LN-EntropyLLM=Llama-3-8B2026.05 | 65.74 | — | — | — | — | — | — | — | |
| SelfCheckGPTBackbone=Llama-3-8B2026.01 | 65.44 | 63.13 | 48.23 | — | — | — | — | — | |
| SelfCheckGPTBackbone=Llama2-13B2026.01 | 65.24 | 63.52 | 54.33 | — | — | — | — | — | |
| LN-EntropyBackbone=Llama2-13B2026.01 | 65.18 | 63.53 | 48.09 | — | — | — | — | — | |
| FActScoreBackbone=Llama2-13B2026.01 | 65.15 | 63.71 | 54.61 | — | — | — | — | — | |
| LN-EntropyBackbone=Llama-3-8B2026.01 | 64.89 | 60.72 | 46.39 | — | — | — | — | — | |
| MINDBackbone=Llama-3-8B2026.01 | 64.54 | 60.89 | 45.13 | — | — | — | — | — | |
| NLILLM Backbone=Gemma-2-9B2025.05 | 64.4 | — | — | — | — | — | — | — | |
| SEPLLM=Gemma-2-9B2026.05 | 64.39 | — | — | — | — | — | — | — | |
| Lexical Sim.Backbone=Llama2-70B2026.01 | 64.37 | 60.92 | 50.86 | — | — | — | — | — | |
| SelfCheckGPTBackbone=Llama2-7B2026.01 | 64.29 | 61.83 | 45.49 | — | — | — | — | — |