Cleft sentence structure analysis on CausalGym cleft
100CompletenessHDMI
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| HDMIModel=Meta-Llama-3-8B-Instruct2026.05 | 100 | 98 | 98.99 | |
| FGSMModel=Meta-Llama-3-8B-Instruct2026.05 | 100 | 100 | 100 | |
| HDMIModel=EleutherAI/pythia-70m2026.05 | 100 | 100 | 100 | |
| HDMIModel=Meta-Llama-3-8B-Instruct2026.05 | 100 | 98 | 98.99 | |
| HDMIModel=EleutherAI/pythia-70m2026.05 | 100 | 100 | 100 | |
| FGSMModel=Meta-Llama-3-8B-Instruct2026.05 | 100 | 100 | 100 | |
| PGDModel=Meta-Llama-3-8B-Instruct2026.05 | 99.46 | 100 | 99.73 | |
| PGDModel=Meta-Llama-3-8B-Instruct2026.05 | 99.46 | 100 | 99.73 | |
| PGDModel=EleutherAI/pythia-70m2026.05 | 95.99 | 100 | 97.96 | |
| PGDModel=EleutherAI/pythia-70m2026.05 | 95.99 | 100 | 97.96 | |
| AlterRepModel=Meta-Llama-3-8B-Instruct2026.05 | 86.03 | 98 | 91.63 | |
| AlterRepModel=Meta-Llama-3-8B-Instruct2026.05 | 86.03 | 98 | 91.63 | |
| FGSMModel=EleutherAI/pythia-70m2026.05 | 79 | 100 | 88.27 | |
| FGSMModel=EleutherAI/pythia-70m2026.05 | 79 | 100 | 88.27 | |
| AlterRepModel=EleutherAI/pythia-70m2026.05 | 73 | 100 | 84.4 | |
| AlterRepModel=EleutherAI/pythia-70m2026.05 | 73 | 100 | 84.4 |