Hirundo Blog company blog

DeepSeek's Hidden Bias: How We Cut It by 76% Without Performance Loss

Showcasing Hirundo’s bias unlearning capabilities on DeepSeek-R1-Distill-Llama-8B, reducing its biases up to 76% compared to original performance.
Points
108
HN comments
119
Submissions
1
Cited by
0 commenters

AI & Machine LearningTech Industry & Business2025

On Hacker News

SubmittedTitle on HNPointsComments
Jan 29, 2025DeepSeek's Hidden Bias: How We Cut It by 76% Without Performance Loss by nicolevin108119

More from Hirundo Blog

All 2 posts →

Related posts from the same era