Twitxy Saved

Charvi Rastogi

@charvvvv_

Researcher in AI & Society @ Google DeepMind
PhD in Machine Learning from CMU
Joined May 2024
343 Following 159 Followers 41 Posts

Latest posts 5

Charvi Rastogi @charvvvv_
New work by @rkdsaakyan showing relevance of diversity of cultural values in safety annotations, along with important meta-analysis of safety datasets in the literature highlighting inconsistent diversification and analytical choices making pluralistic safety research harder
Arkadiy Saakyan @rkdsaakyan Jun 2
Excited to share #ICML2026 paper from my internship @GoogleDeepMind!
AI models are deployed globally, but safety datasets are largely geographically homogenous. What is the impact of culture on AI safety ratings? Is there impact beyond standard demographics?
View quoted post on X
Charvi Rastogi @charvvvv_
Presenting this tomorrow (Friday 11AM-2PM PST) with @laroyo & @ImPushMish! Come chat with us about Pluralistic AI, safety measurement and AI evaluations, and if you work on alignment let's discuss the DIVE dataset and using it for designing pluralistically safe and inclusive AI.
📜 💡**New paper & dataset** Our paper [to be presented as a spotlight at NeurIPS’25] presents DIVE, a novel dataset for aligning GenAI models to pluralistic viewpoints. DIVE and accompanying experiments advocate for a shift in both: alignment strategies and safety evaluations.
View quoted post on X
Charvi Rastogi @charvvvv_
📜 💡**New paper & dataset** Our paper [to be presented as a spotlight at NeurIPS’25] presents DIVE, a novel dataset for aligning GenAI models to pluralistic viewpoints. DIVE and accompanying experiments advocate for a shift in both: alignment strategies and safety evaluations.