diff options
Diffstat (limited to 'research/flossing/late_perturb_robustness/smoke_plots')
| -rw-r--r-- | research/flossing/late_perturb_robustness/smoke_plots/late_perturb_robustness_combined.csv | 5 |
1 files changed, 5 insertions, 0 deletions
diff --git a/research/flossing/late_perturb_robustness/smoke_plots/late_perturb_robustness_combined.csv b/research/flossing/late_perturb_robustness/smoke_plots/late_perturb_robustness_combined.csv new file mode 100644 index 0000000..ec4de26 --- /dev/null +++ b/research/flossing/late_perturb_robustness/smoke_plots/late_perturb_robustness_combined.csv @@ -0,0 +1,5 @@ +label,perturb_after,sigma,n_samples,rollouts,ckpt_root,ckpt_name,perturb,noise_distribution,mean_rollout_exact,mean_rollout_token_acc,pass_at_k,all_k,correct_count_mean,correct_count_std,zero_frac,full_frac,clean_acc,retain_mean_on_clean_success,allK_on_clean_success,rescue_mean_on_clean_fail,passK_on_clean_fail
+smoke_baseline,0,0.0,16,2,/home/yurenh2/rrm/trm/checkpoints/Sudoku-extreme-1k-aug-1000-ACT-torch/pretrain_mlp_t_sudoku_official_gbs768_repro,step_58590,both,gaussian,0.875,0.9614197611808777,0.875,0.875,1.75,0.6614378094673157,0.125,0.875,0.875,1.0,1.0,0.0,0.0
+smoke_baseline,0,0.01,16,2,/home/yurenh2/rrm/trm/checkpoints/Sudoku-extreme-1k-aug-1000-ACT-torch/pretrain_mlp_t_sudoku_official_gbs768_repro,step_58590,both,gaussian,0.9375,0.9834104776382446,1.0,0.875,1.875,0.33071890473365784,0.0,0.875,0.875,0.9642857313156128,0.9285714030265808,0.75,1.0
+smoke_baseline,8,0.0,16,2,/home/yurenh2/rrm/trm/checkpoints/Sudoku-extreme-1k-aug-1000-ACT-torch/pretrain_mlp_t_sudoku_official_gbs768_repro,step_58590,both,gaussian,0.875,0.9614197611808777,0.875,0.875,1.75,0.6614378094673157,0.125,0.875,0.875,1.0,1.0,0.0,0.0
+smoke_baseline,8,0.01,16,2,/home/yurenh2/rrm/trm/checkpoints/Sudoku-extreme-1k-aug-1000-ACT-torch/pretrain_mlp_t_sudoku_official_gbs768_repro,step_58590,both,gaussian,0.90625,0.96875,0.9375,0.875,1.8125,0.5266343355178833,0.0625,0.875,0.875,1.0,1.0,0.25,0.5
|
