Srishti Gureja

Independent

Papers on Lattice

Total citations

Topics

Publication activitypapers/week, last 8 weeks

Research focus

Constitutional AI & AI Ethics (1)Red-Teaming & Adversarial Robustness (1)

Frequent co-authors

Kajetan Dymkiewicz (1)Tim Farrelly (1)Adam Práda (1)Ishaan Panigrahi (1)

Papers (1)

Sep 28, 2026

1d ago·also Affiliation:, Independent, KCL

Don't Inoculate Everything: Stratified Inoculation Prompting Narrows Backdoor Triggers and Preserves Desired Traits

Standard inoculation prompting often cripples capabilities and fails to contain bad behaviors, but diversifying prompt contexts across just a tiny fraction of clean data effectively cordons off emergent misalignment.

Kajetan Dymkiewicz, Tim Farrelly, Adam Práda +7

Constitutional AI & AI Ethics Red-Teaming & Adversarial Robustness

Search

Srishti Gureja

Publication activitypapers/week, last 8 weeks

Research focus

Frequent co-authors

Papers (1)