Karthikeyan Shanmugam

Researcher@Google Deepmind·印度·OpenReview

研究方向

Online Learning and Bandits · Causal Inference · Foundation Models · Diffusion Models · Reinforcement Learning · Graphical Models · Information Theory

6.0

Does Safety Training of LLMs Generalize to Semantically Related Natural Prompts?

ICLR 2025Poster

4.8

TreeTop: Topology-Aware Fine-Tuning for LLM Conversation Tree Understanding

合作者 (20)

Dheeraj Mysore Nagaraj

2 篇

Karthikeyan Shanmugam

Efficient Approximate Posterior Sampling with Annealed Langevin Monte Carlo

Robust Reward Modeling via Causal Rubrics

ROPES: Robotic Pose Estimation via Score-based Causal Representation Learning

Fine-Tuning Diffusion Models via Intermediate Distribution Shaping

Path-specific effects for pulse-oximetry guided decisions in critical care

Glauber Generative Model: Discrete Diffusion Models via Binary Classification

Does Safety Training of LLMs Generalize to Semantically Related Natural Prompts?

TreeTop: Topology-Aware Fine-Tuning for LLM Conversation Tree Understanding