rlxf.ai is a blog and research site maintained by Shixiang Shane Gu, an AI researcher at Google DeepMind, currently working on Gemini and Omni Thinking. Previously he led the Multilinguality team for Gemini Post-Training (Gemini 1.5–2.5 production models). Shane is recognized for pioneering contributions across generative modeling, reinforcement learning, and LLM reasoning — including co-inventing Gumbel-Softmax, Zero-Shot Chain-of-Thought prompting ("Let's think step by step"), and showing that LLMs can self-improve. This site collects his thoughts, writing, and research updates.
Latest Posts
- No posts found.
Selected Publications
MuProp: Unbiased Backpropagation for Stochastic Neural Networks
Shixiang Gu, Sergey Levine, Ilya Sutskever, Andriy Mnih
ICLR 2016
Deep RL for Robotic Manipulation with Asynchronous Off-Policy Updates
Shixiang Gu, Ethan Holly, Timothy Lillicrap, Sergey Levine
ICRA 2017
Blocks Assemble! Learning to Assemble with Large-Scale Structured RL
Seyed Kamyar Seyed Ghasemipour, Daniel Freeman, Byron David, Shixiang Shane Gu, Satoshi Kataoka, Igor Mordatch
ICML 2022
Continuous Deep Q-Learning with Model-based Acceleration
Shixiang Gu, Timothy Lillicrap, Ilya Sutskever, Sergey Levine
ICML 2016
A Divergence Minimization Perspective on Imitation Learning Methods
Best Paper AwardSeyed Kamyar Seyed Ghasemipour, Richard Zemel, Shixiang Gu
CoRL 2019
Large Language Models are Zero-Shot Reasoners
Takeshi Kojima, Shixiang Shane Gu, Machel Reid, Yutaka Matsuo, Yusuke Iwasawa
NeurIPS 2022
Large Language Models Can Self-Improve
Jiaxin Huang, Shixiang Shane Gu, Le Hou, Yuexin Wu, Xuezhi Wang, Hongkun Yu, Jiawei Han
EMNLP 2023
Video Models are Zero-Shot Reasoners
Thaddäus Wiedemer, Yuxuan Li, Paul Vicol, Shixiang Shane Gu, Nick Matarese, Kevin Swersky, Been Kim, Priyank Jaini, Robert Geirhos
arXiv 2025
Mind's Eye: Grounded Language Model Reasoning through Simulation
Ruibo Liu, Jason Wei, Shixiang Shane Gu, Te-Yen Wu, Soroush Vosoughi, Claire Cui, Denny Zhou, Andrew M. Dai
ICLR 2023