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

Bias Evaluation on Honest

11.7Honest Score

Llama 3.1 8B - LFT w. SH-Dgender (BaseCDA)

Updated 1mo ago

Evaluation Results

Method	Links
Llama 3.1 8B - LFT w. SH-Dgender (BaseCDA) 2025.12		11.7
Llama 3.1 8B - LFT w. SH-Dgender (GC-CDA) 2025.12		12.1
Llama 3.1 8B - LFT w. SH (baseline 2) 2025.12		12.4
Qwen 3 0.6B - LFT w. SH-N (baseline 3) 2025.12		13.2
Llama 3.2 1B - LFT w. SH (baseline 2) 2025.12		14.6
Llama 3.1 8B - LFT w. SH-N (baseline 3) 2025.12		14.6
Llama 3.1 8B - Pretrained model (baseline 1) 2025.12		14.7
Llama 3.2 1B - LFT w. SH-Dgender (BaseCDA) 2025.12		14.8
Qwen 3 0.6B - Pretrained model (baseline 1) 2025.12		15.8
Llama 3.2 1B - LFT w. SH-Dgender (GC-CDA) 2025.12		15.9
Qwen 3 0.6B - LFT w. SH-Dgender(BaseCDA) 2025.12		17.8
Qwen 3 0.6B - LFT w. SH (baseline 2) 2025.12		18.2
Qwen 3 0.6B - LFT w. SH-Dgender (GC-CDA) 2025.12		19.7
Llama 3.2 1B - Pretrained model (baseline 1) 2025.12		20.8
Llama 3.2 1B - LFT w. SH-N (baseline 3) 2025.12		21.4