Divya Chaudhary

Research focus

Constitutional AI & AI Ethics (2)Scalable Oversight & Alignment Theory (2)Reasoning & Chain-of-Thought (1)RLHF & Preference Learning (1)

Frequent co-authors

Subramanyam Sahoo (2)Aman Chadha (2)Vinija Jain (2)

Papers (2)

Mar 10, 2026

Subramanyam Sahoo +36d ago

The Reasoning Trap -- Logical Reasoning as a Mechanistic Pathway to Situational Awareness

LLM reasoning research is inadvertently paving a dangerous path towards AI situational awareness and strategic deception, demanding a re-evaluation of current safety measures.

Subramanyam Sahoo, Aman Chadha, Vinija Jain +1

Constitutional AI & AI Ethics Reasoning & Chain-of-Thought Scalable Oversight & Alignment Theory

Mar 6, 2026

Subramanyam Sahoo +31w ago

SAHOO: Safeguarded Alignment for High-Order Optimization Objectives in Recursive Self-Improvement

Recursive self-improvement can boost performance by 18% in code and 17% in reasoning, but only if you can keep it from going off the rails – SAHOO provides the guardrails.

Subramanyam Sahoo, Aman Chadha, Vinija Jain +1

Constitutional AI & AI Ethics RLHF & Preference Learning Scalable Oversight & Alignment Theory

Search

Divya Chaudhary

Publication activitypapers/week, last 8 weeks

Research focus

Frequent co-authors

Papers (2)