Udari Madhushani Sehwag
4 papers · 2025–2026 · 2 venues · across top CS/AI venues
Authors are matched by name, so researchers who share a name ("Yang Liu", "Wei Wang") appear as a single profile and their paper counts, venues and achievements are combined. How identity is handled.
Achievements
Jump to papers ↓
🌍
Conference Polyglot · Bronze
🌉
Interdisciplinary Bridge · Bronze
Top co-authors
Research topics
Keywords
large language model
(3)
autoregressive generation
(1)
harmful content
(1)
reinforcement learning from human feedback
(1)
model alignment
(1)
kl regularization
(1)
backdoor attack
(1)
reward model
(1)
test-time alignment
(1)
large language model alignment
(1)
adversarial fine-tuning
(1)
controlled decoding
(1)
multi-agent system
(1)
unsafe content
(1)
reinforcement learning human feedback
(1)
preference alignment
(1)
safety refusal
(1)
Papers
AdvBDGen: A Robust Framework for Generating Adaptive and Stealthy Backdoors in LLM Alignment
AAAI 2026
Collab: Controlled Decoding using Mixture of Agents for LLM Alignment
ICLR 2025
GenARM: Reward Guided Generation with Autoregressive Reward Model for Test-Time Alignment
ICLR 2025
SORRY-Bench: Systematically Evaluating Large Language Model Safety Refusal
ICLR 2025