Skip to content
View jyyang26's full-sized avatar

Block or report jyyang26

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
jyyang26/README.md

Hi there, I'm Junyao 👋

杨竣尧 · National University of Singapore · Research Intern @ Tencent Hy Frontier, Singapore

Homepage Google Scholar Email LinkedIn X Total GitHub Stars


About Me

I am a graduate student at the School of Computing, National University of Singapore, specializing in Artificial Intelligence, and currently a Research Intern at Tencent Hy Frontier Lab, Singapore, where I work on Agentic RL, long-horizon terminus agents, and RL stability.

Before that, I was a Research Intern at Shanghai AI Lab and a Research Assistant at ZeroNLP, and I received my B.S. in Computer Science (with honors) from South China University of Technology.

Research Interests

  • 🤖 Agentic AI — long-horizon self-evolving terminal agents
  • 🧠 Large Language Models — reasoning, model merging, capability transfer
  • 🎯 Reinforcement Learning — training stability, asynchronous RL, reward design
  • 🔍 Explainable AI — internal mechanisms and attribution of model behavior

Selected Publications

Venue Paper
Preprint Terminal-RSI: Recursive Self-Improving Terminal Agents via Recursive Self-Rewrite and Policy-Refine
Tech Report T1: Terminal Agent Reinforcement Learning for Long-Horizon Tasks
ACL 2026 ReasonAny: Incorporating Reasoning Capability to Any Model via Simple and Effective Model Merging · code
AAAI 2026 RCP-Merging: Merging Long Chain-of-Thought Models with Domain-Specific Models · code
ACL 2025 PrivacyRestore: Privacy-Preserving Inference in LLMs via Privacy Removal and Restoration · code
EMNLP 2025 RewardDS: Privacy-Preserving Fine-Tuning for LLMs via Reward Driven Data Synthesis · code
Tech Report AgentDoG: A Diagnostic Guardrail Framework for AI Agent Safety and Security · code
Preprint RSR · SWR · Long-Horizon-Terminal-Bench · RST · Harness Handbook · Entropy-Gradient Inversion

📄 Full list on my homepage · CV: EN / 中文


Feel free to reach out if you'd like to chat about agents, reasoning, or RL. 📫 junyaoyang@u.nus.edu

Pinned Loading

  1. Terminal-RSI Terminal-RSI Public

    Official Codebase of "Terminal-RSI: Recursive Self-Improving Terminal Agents via Recursive Self-Rewrite and Policy-Refine"

    Python 82 18

  2. AI45Lab/AgentDoG AI45Lab/AgentDoG Public

    A Diagnostic Guardrail Framework for AI Agent Safety and Security

    Python 710 38

  3. zli12321/LHTB zli12321/LHTB Public

    Long Horizon Terminal Benchmark with Dense Reward Grading

    Python 570 35

  4. SAT SAT Public

    Staleness-Adaptive Trust Region (SAT), a directional, token-adaptive clipping method for stabilizing asynchronous reinforcement learning of large language models.

    Python 24 1

  5. ReasonAny ReasonAny Public

    Official Code of "ReasonAny: Incorporating Reasoning Capability to Any Model via Simple and Effective Model Merging".

    Python 9

  6. ZeroNLP/RCP-Merging ZeroNLP/RCP-Merging Public

    [AAAI2026]: Official Code of "RCP-Merging: Merging Long Chain-of-Thought Models with Domain-Specific Models by Considering Reasoning Capability as Prior"

    Python 10 1