Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Prompt Injection on GSM8k

3.79ASR@3

G-Safeguard

3.16847.364211.5615.7558Aug 11, 2025Sep 2, 2025Sep 24, 2025Oct 16, 2025Nov 7, 2025Nov 29, 2025Dec 21, 2025
Updated 1mo ago

Evaluation Results

MethodLinks
2025.08
3.790.9911
2025.08
4.40.9822
2025.12
60.9778
2025.08
6.10.9822
2025.08
6.440.7556
2025.12
6.670.9911
2025.08
6.780.6844
2025.08
6.780.7022
2025.08
7.80.9911
2025.08
7.80.6978
2025.08
8.140.5467
2025.08
8.140.6844
2025.08
8.470.6756
2025.08
8.470.6933
2025.08
8.470.5333
2025.08
8.470.5955
2025.12
8.670.9778
2025.08
8.790.6222
2025.08
8.840.5156
2025.12
8.890.6296
2025.12
9.330.9911
2025.12
9.670.9733
2025.12
9.670.7111
2025.12
9.670.9911
2025.12
9.830.6701
2025.08
9.83-
2025.12
10.180.6897
2025.08
10.2-
2025.12
10.290.5962
2025.12
10.330.7067
2025.12
10.330.9867
2025.12
10.510.6881
2025.08
10.510.5956
2025.08
10.510.6978
2025.08
10.510.6978
2025.12
10.670.6533
2025.12
10.910.6655
2025.12
11.670.5289
2025.12
11.670.5467
2025.12
11.670.9733
2025.08
12.89-
2025.12
13.080.641
2025.12
13.67-
2025.12
13.67-
2025.12
140.5599
2025.12
14.330.7422
2025.08
14.48-
2025.08
14.910.4622
2025.12
15-
2025.12
15.670.4844
2025.12
16.140.6257
2025.12
19.33-