Binary ill-posedness detection on SituatedQA
90.4AccuracyH0 topology
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| H0 topologyModel=Llama-3.1-8B-Instruct2026.06 | 90.4 | 90.4 | |
| H0 topologyModel=Mistral-7B-Instruct-v0.32026.06 | 88.4 | 88.4 | |
| H0 topologyModel=Gemma-7B-it2026.06 | 86.6 | 86.6 | |
| AENModel=Mistral-7B-Instruct-v0.32026.06 | 85.5 | 85.5 | |
| AENModel=Llama-3.1-8B-Instruct2026.06 | 80.2 | 80.2 | |
| AENModel=Gemma-7B-it2026.06 | 73.9 | 73.9 | |
| InfoGainModel=Mistral-7B-Instruct-v0.32026.06 | 61.6 | 59.5 | |
| InfoGainModel=Gemma-7B-it2026.06 | 57.7 | 57.6 | |
| Zero-shotModel=Llama-3.1-8B-Instruct2026.06 | 57.1 | 54.6 | |
| Few-shot CoTModel=Mistral-7B-Instruct-v0.32026.06 | 56.4 | 56.3 | |
| InfoGainModel=Llama-3.1-8B-Instruct2026.06 | 55.2 | 46.2 | |
| Few-shotModel=Llama-3.1-8B-Instruct2026.06 | 54.9 | 36.3 | |
| Few-shotModel=Mistral-7B-Instruct-v0.32026.06 | 54.4 | 43.1 | |
| Zero-shotModel=Mistral-7B-Instruct-v0.32026.06 | 54.3 | 46.1 | |
| Few-shot CoTModel=Llama-3.1-8B-Instruct2026.06 | 52.9 | 52.7 | |
| Zero-shotModel=Gemma-7B-it2026.06 | 51.5 | 42.5 | |
| Few-shot CoTModel=Gemma-7B-it2026.06 | 48.6 | 47.9 | |
| Few-shotModel=Gemma-7B-it2026.06 | 47.4 | 32.6 |