Counterfactual Subset-Utility Evaluation on CommonsenseQA
0.26Spearman Correlation ($
ho$)GRASP
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GRASPEvaluation Protocol=Linear Datamodeling Score (LDS), Model Architecture=50M-parameter causal decoder-only Transformer2026.06 | 0.26 | 87 | |
| TracInEvaluation Protocol=Linear Datamodeling Score (LDS), Model Architecture=50M-parameter causal decoder-only Transformer2026.06 | 0.163 | 68 | |
| TRAKEvaluation Protocol=Linear Datamodeling Score (LDS), Model Architecture=50M-parameter causal decoder-only Transformer2026.06 | -0.062 | 49.5 | |
| BM25Evaluation Protocol=Linear Datamodeling Score (LDS), Model Architecture=50M-parameter causal decoder-only Transformer2026.06 | -0.089 | 51 |