Hallucination Detection on SciQ
64.8EigenWD M1 ScoreEigenWD M1
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| EigenWD M1Black model=deepseek-chat, Teacher model=Llama-2-7b, Number of samples=10, Teacher forcing=true2026.03 | 64.8 | — | — | — | — | — | |
| EigenWD M1Black model=deepseek-chat, Teacher model=Llama-3.1-8b, Number of samples=10, Teacher forcing=true2026.03 | 63.92 | — | — | — | — | — | |
| EigenWD M1Black model=chatgpt-4o-mini, Teacher model=Llama-3.1-8b, Number of samples=10, Teacher forcing=true2026.03 | 62.51 | — | — | — | — | — | |
| EigenWD M1Black model=chatgpt-4o-mini, Teacher model=Qwen3-8B, Number of samples=10, Teacher forcing=true2026.03 | 62.28 | — | — | — | — | — | |
| EigenWD M1Black model=deepseek-chat, Teacher model=Qwen3-8B, Number of samples=10, Teacher forcing=true2026.03 | 61.02 | — | — | — | — | — | |
| EigenWD M1Black model=chatgpt-4o-mini, Teacher model=Llama-2-7b, Number of samples=10, Teacher forcing=true2026.03 | 60.32 | — | — | — | — | — | |
| DSEBlack model=chatgpt-4o-mini, Teacher model=Llama-2-7b, Number of samples=10, Teacher forcing=true2026.03 | — | — | — | — | — | 61.6 | |
| DSEBlack model=chatgpt-4o-mini, Teacher model=Llama-3.1-8b, Number of samples=10, Teacher forcing=true2026.03 | — | — | — | — | — | 61.6 | |
| DSEBlack model=chatgpt-4o-mini, Teacher model=Qwen3-8B, Number of samples=10, Teacher forcing=true2026.03 | — | — | — | — | — | 61.6 | |
| DSEBlack model=deepseek-chat, Teacher model=Llama-2-7b, Number of samples=10, Teacher forcing=true2026.03 | — | — | — | — | — | 59.78 | |
| DSEBlack model=deepseek-chat, Teacher model=Llama-3.1-8b, Number of samples=10, Teacher forcing=true2026.03 | — | — | — | — | — | 59.78 | |
| DSEBlack model=deepseek-chat, Teacher model=Qwen3-8B, Number of samples=10, Teacher forcing=true2026.03 | — | — | — | — | — | 59.78 | |
| ERBlack model=chatgpt-4o-mini, Teacher model=Llama-2-7b, Number of samples=10, Teacher forcing=true2026.03 | — | 63.82 | — | — | — | — | |
| ERBlack model=chatgpt-4o-mini, Teacher model=Llama-3.1-8b, Number of samples=10, Teacher forcing=true2026.03 | — | 61.08 | — | — | — | — | |
| ERBlack model=chatgpt-4o-mini, Teacher model=Qwen3-8B, Number of samples=10, Teacher forcing=true2026.03 | — | 61.96 | — | — | — | — | |
| ERBlack model=deepseek-chat, Teacher model=Llama-2-7b, Number of samples=10, Teacher forcing=true2026.03 | — | 63.98 | — | — | — | — | |
| ERBlack model=deepseek-chat, Teacher model=Llama-3.1-8b, Number of samples=10, Teacher forcing=true2026.03 | — | 59.88 | — | — | — | — | |
| ERBlack model=deepseek-chat, Teacher model=Qwen3-8B, Number of samples=10, Teacher forcing=true2026.03 | — | 60.9 | — | — | — | — | |
| ESBlack model=chatgpt-4o-mini, Teacher model=Llama-2-7b, Number of samples=10, Teacher forcing=true2026.03 | — | — | 62.52 | — | — | — | |
| ESBlack model=chatgpt-4o-mini, Teacher model=Llama-3.1-8b, Number of samples=10, Teacher forcing=true2026.03 | — | — | 61.48 | — | — | — | |
| ESBlack model=chatgpt-4o-mini, Teacher model=Qwen3-8B, Number of samples=10, Teacher forcing=true2026.03 | — | — | 61.54 | — | — | — | |
| ESBlack model=deepseek-chat, Teacher model=Llama-2-7b, Number of samples=10, Teacher forcing=true2026.03 | — | — | 57.84 | — | — | — | |
| ESBlack model=deepseek-chat, Teacher model=Llama-3.1-8b, Number of samples=10, Teacher forcing=true2026.03 | — | — | 63.14 | — | — | — | |
| ESBlack model=deepseek-chat, Teacher model=Qwen3-8B, Number of samples=10, Teacher forcing=true2026.03 | — | — | 65.34 | — | — | — | |
| LNEBlack model=chatgpt-4o-mini, Teacher model=Llama-2-7b, Number of samples=10, Teacher forcing=true2026.03 | — | — | — | 62.23 | — | — | |
| LNEBlack model=chatgpt-4o-mini, Teacher model=Llama-3.1-8b, Number of samples=10, Teacher forcing=true2026.03 | — | — | — | 60.38 | — | — | |
| LNEBlack model=chatgpt-4o-mini, Teacher model=Qwen3-8B, Number of samples=10, Teacher forcing=true2026.03 | — | — | — | 65.37 | — | — | |
| LNEBlack model=deepseek-chat, Teacher model=Llama-2-7b, Number of samples=10, Teacher forcing=true2026.03 | — | — | — | 59.16 | — | — | |
| LNEBlack model=deepseek-chat, Teacher model=Llama-3.1-8b, Number of samples=10, Teacher forcing=true2026.03 | — | — | — | 57.82 | — | — | |
| LNEBlack model=deepseek-chat, Teacher model=Qwen3-8B, Number of samples=10, Teacher forcing=true2026.03 | — | — | — | 59.27 | — | — | |
| LSBlack model=chatgpt-4o-mini, Teacher model=Llama-2-7b, Number of samples=10, Teacher forcing=true2026.03 | — | — | — | — | 61.45 | — | |
| LSBlack model=chatgpt-4o-mini, Teacher model=Llama-3.1-8b, Number of samples=10, Teacher forcing=true2026.03 | — | — | — | — | 61.45 | — | |
| LSBlack model=chatgpt-4o-mini, Teacher model=Qwen3-8B, Number of samples=10, Teacher forcing=true2026.03 | — | — | — | — | 61.45 | — | |
| LSBlack model=deepseek-chat, Teacher model=Llama-2-7b, Number of samples=10, Teacher forcing=true2026.03 | — | — | — | — | 49.21 | — | |
| LSBlack model=deepseek-chat, Teacher model=Llama-3.1-8b, Number of samples=10, Teacher forcing=true2026.03 | — | — | — | — | 49.21 | — | |
| LSBlack model=deepseek-chat, Teacher model=Qwen3-8B, Number of samples=10, Teacher forcing=true2026.03 | — | — | — | — | 49.21 | — |