Skip to content
View XavierZhang2002's full-sized avatar
8 followers · 5 following

Achievements

Achievement: Starstruck

Achievements

Achievement: Starstruck

Block or report XavierZhang2002

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. Markdown is supported. 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
XavierZhang2002/README.md

Hi there, I'm Zhenliang Zhang 👋

I'm a Master's student at Peking University, advised by Prof. Xiaojun Wan. My research focuses on LLM post-training for real-world agentic scenarios, including long-text understanding and financial analysis. Previously, I studied LLM reliability through a progression from causal analysis of hallucinations (CIKM 2025), through reliable hallucination detection (ACL 2025), to representation-level intervention and controlled generation (AAAI 2026).

Currently, I am an intern on the Ling Team at Ant Group, working on Agent Instruction Following (IF) post-training. Previously, I was a research intern at Tencent WeChat, where I developed SCOUT (ICML 2026), an agentic active information-foraging paradigm for million-token long-text understanding.

Pinned Loading

  1. ICR_Probe ICR_Probe Public

    Code repository of "ICR Probe: Tracking Hidden State Dynamics for Reliable Hallucination Detection in LLMs" (ACL 2025).

    Jupyter Notebook 23 5

  2. agentic-rl-analysis agentic-rl-analysis Public

    A systematic survey of LLM post-training algorithms, industrial practices, and frontier directions.

    7

  3. BID BID Public

    Resources for the paper ​​"Exploring Causal Effect of Social Bias on Faithfulness Hallucinations in Large Language Models" (CIKM 2025).

    2

  4. agent-state-management agent-state-management Public

    A source-level deep dive into how Claude Code and OpenClaw manage agent state.

  5. scout-page scout-page Public

    Project page for SCOUT: Active Information Foraging for Long-Text Understanding with Decoupled Epistemic States (ICML 2026)

    JavaScript

  6. scout-open scout-open Public

    Open-source implementation of SCOUT: Active Information Foraging for Long-Text Understanding with Decoupled Epistemic States (ICML 2026).

    TypeScript 7