gss_subord_pp on CausalGym
100CompletenessHDMI
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| HDMIModel=Meta-Llama-3-8B-Instruct2026.05 | 100 | 89.69 | 94.56 | |
| FGSMModel=Meta-Llama-3-8B-Instruct2026.05 | 100 | 93.82 | 96.81 | |
| HDMIModel=EleutherAI/pythia-70m2026.05 | 99.84 | 73.71 | 84.81 | |
| PGDModel=Meta-Llama-3-8B-Instruct2026.05 | 99.5 | 52.98 | 69.14 | |
| AlterRepModel=Meta-Llama-3-8B-Instruct2026.05 | 97 | 89.38 | 93.03 | |
| AlterRepModel=EleutherAI/pythia-70m2026.05 | 68.96 | 70.51 | 69.72 | |
| PGDModel=EleutherAI/pythia-70m2026.05 | 30.15 | 66.28 | 41.44 | |
| FGSMModel=EleutherAI/pythia-70m2026.05 | 13.33 | 97.99 | 23.47 |