Hallucination Detection on SQuAD 2 (test)
92.7AUROCTriLens
Evaluation Results
| Method | Links | |
|---|---|---|
| TriLensLLM=Gemma-2-9B2026.05 | 92.7 | |
| TriLensLLM=Llama-3-8B2026.05 | 91.69 | |
| TriLensLLM=Qwen2.5-7B2026.05 | 90.61 | |
| ICR ProbeLLM=Gemma-2-9B2026.05 | 81.42 | |
| ICR ProbeLLM=Llama-3-8B2026.05 | 75.68 | |
| SAPLMALLM=Gemma-2-9B2026.05 | 74.09 | |
| ICR ProbeLLM=Qwen2.5-7B2026.05 | 73.82 | |
| SEPLLM=Llama-3-8B2026.05 | 72.74 | |
| SAPLMALLM=Llama-3-8B2026.05 | 70.43 | |
| SAPLMALLM=Qwen2.5-7B2026.05 | 70.16 | |
| LN-EntropyLLM=Gemma-2-9B2026.05 | 68.18 | |
| SEPLLM=Gemma-2-9B2026.05 | 66.27 | |
| LN-EntropyLLM=Qwen2.5-7B2026.05 | 65.64 | |
| PPLLLM=Llama-3-8B2026.05 | 64.72 | |
| SEPLLM=Qwen2.5-7B2026.05 | 64.18 | |
| LN-EntropyLLM=Llama-3-8B2026.05 | 62.49 | |
| LLM-CheckLLM=Qwen2.5-7B2026.05 | 56.39 | |
| LLM-CheckLLM=Gemma-2-9B2026.05 | 55.17 | |
| LLM-CheckLLM=Llama-3-8B2026.05 | 53.56 | |
| PPLLLM=Gemma-2-9B2026.05 | 53.48 | |
| PPLLLM=Qwen2.5-7B2026.05 | 52.78 |