Skip to content

Bias Eval TR: BBQ-TR — Gender / Ethnicity / Sect / Age / Socioeconomic Probe + Mitigation

BBQ (Bias Benchmark for QA, Parrish 2022) TR adaptation: gender, ethnicity (Turkish/Kurdish/Arab/Armenian), sect (Sunni/Alevi), age, socioeconomic, physical appearance — 9 categories bias probe. 1200 ambiguous question pairs. Cookbook's mitigation recipe: balanced SFT data + DPO bias-rejection examples.

Şükrü Yusuf KAYA
28 min read
Advanced
Bias Eval TR: BBQ-TR — Cinsiyet / Etnik / Mezhep / Yaş / SES Probe + Mitigation
✅ Teslim
  1. BBQ-TR (cookbook reference) ile model bias score ölç. 2) DPO bias-rejection pair üret. 3) Sonraki ders: 18.7 — Red-Teaming Lab.

Yorumlar & Soru-Cevap

(0)
Yorum yazmak için giriş yap.
Yorumlar yükleniyor...

Related Content