Hallucination prediction on Explain (+ domain)
64.87AccuracyConf + Probe (SCAO)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Conf + Probe (SCAO)Model=70B2025.09 | 64.87 | 9.56 | |
| Conf + Probe (SCAO)Model=LLaMA3-70B-Instruct2025.09 | 64.87 | 9.56 | |
| Conf + ProbeModel=8B2025.09 | 64.78 | 3.57 | |
| Conf + ProbeModel=LLaMA3-8B-Instruct2025.09 | 64.78 | 3.57 | |
| Conf + Probe (SCAO)Model=8B2025.09 | 63.66 | 2.45 | |
| Conf + Probe (SCAO)Model=LLaMA3-8B-Instruct2025.09 | 63.66 | 2.45 | |
| Probe_dnnModel=8B2025.09 | 63.31 | 2.1 | |
| ProbednnModel=LLaMA3-8B-Instruct2025.09 | 63.31 | 2.1 | |
| Conf + ProbeModel=70B2025.09 | 63.1 | 7.79 | |
| Conf + ProbeModel=LLaMA3-70B-Instruct2025.09 | 63.1 | 7.79 | |
| Probe_dnnModel=70B2025.09 | 62.65 | 7.34 | |
| ProbednnModel=LLaMA3-70B-Instruct2025.09 | 62.65 | 7.34 | |
| Conf (SCAO)Model=70B2025.09 | 60.78 | — | |
| Conf (SCAO)Model=LLaMA3-70B-Instruct2025.09 | 60.78 | — | |
| ConfModel=8B2025.09 | 60.29 | — | |
| ConfModel=LLaMA3-8B-Instruct2025.09 | 60.29 | — | |
| Conf (SCAO)Model=8B2025.09 | 59.57 | — | |
| Conf (SCAO)Model=LLaMA3-8B-Instruct2025.09 | 59.57 | — | |
| ConfModel=70B2025.09 | 53.31 | — | |
| ConfModel=LLaMA3-70B-Instruct2025.09 | 53.31 | — |