🏆 Bias Leaderboard for Gendered Prompts

*GG = Gender Gap
Leaderboard for cover letters that were generated with gendered prompts.

# Model Manual annotation GG Masc-Gendered GG Fem-Gendered Gender Shift Corpus size (number of letter) Date
1 bloom-560m 1.15 - 65.09 4451 2025-05-29 08:26:21
2 xglm-2 7.05 - 32.79 4449 2025-05-29 08:26:21
3 mistral-7b-v0.3 - 7.73 45.18 3818 2025-05-29 08:26:21
4 croissantbase 9.07 - 57.71 4181 2025-05-29 08:26:21
5 llama-3.2-3b 10.05 - 40.54 4711 2025-05-29 08:26:21
6 gemma-2-2b 10.39 - 56.71 4893 2025-05-29 08:26:21
7 croissant-chat* 11.44 - 66.32 4931 2025-05-29 08:26:21
8 bloom-3b 17.23 - 71.82 4396 2025-05-29 08:26:21
9 vigogne-2-7b 18.4 - 64.69 4384 2025-05-29 08:26:21
10 bloom-7b 19.93 - 65.78 4161 2025-05-29 08:26:21
11 gpt2-fr 21.81 - 60.35 4852 2025-05-29 08:26:21
12 llama-3.2-3b-it* 25.47 - 83.37 4967 2025-05-29 08:26:21
13 gemma-2-2b-it* 28.88 - 46.59 4595 2025-05-29 08:26:21

*Sorted by GenderGap