mcptest-user

#!/usr/bin/env bash
set -euo pipefail

# Bootstrap script to create LJT-Homepage from academicpages template and personalize with Junteng Liu's info
# Requirements: curl, git, and a GitHub token with repo scope in $GITHUB_TOKEN. Optionally gh CLI for convenience.

if [[ -z "${GITHUB_TOKEN:-}" ]]; then
  echo "Error: Please export GITHUB_TOKEN with 'repo' scope." >&2
  exit 1
fi

mcptest-user

Junteng Liu\n\n## Contact Information\n- Email: jliugi@connect.ust.hk\n- GitHub: Vicent0205\n- Google Scholar: Profile\n- X (Twitter): @junteng88716710\n\n## Education\n- Ph.D. in Computer Science (2024-Present), Hong Kong University of Science and Technology\n- B.Eng. (2020-2024), Shanghai Jiao Tong University\n\n## Research Interests\n- Natural Language Processing\n- Machine Learning\n- LLM Reasoning and Reinforcement Learning\n- Hallucination in Vision-Language Models (VLM)\n- LLM Truthfulness and Interpretability\n\n## Advisors\n- Professor Junxian He (PhD supervisor at HKUST NLP Group)\n\n## Research Experience\n- Research Intern, MINIMAX (February 2025 - Present)\n- Research Intern, Tencent WXG (June 2024 - September 2024)\n - Advisor: Zifei Shan\n- Research Intern, Shanghai AI Lab (June 2023 - December 2023)\n - Advisor: Prof. Yu C

mcptest-user

Configuration Files for Academic Pages Template\n\n## _config.yml\n\n```yaml\ntitle: Junteng Liu\nemail: jliugi@connect.ust.hk\ndescription: >\n PhD Candidate in Computer Science at HKUST NLP Group\n\n# Social Media Links\ngithub_username: Vicent0205\ngithub_url: https://github.com/Vicent0205\ntwitter_username: junteng88716710\ntwitter_url: https://twitter.com/junteng88716710\ngoogle_scholar_url: https://scholar.google.com/citations?hl=en&user=tbK9jl4AAAAJ&view_op=list_works&sortby=pubdate\n\n# Build settings\nmarkdown: kramdown\ntheme: minimal\nplugins:\n - jekyll-feed\n\n# Collections\ncollections:\n publications:\n output: true\n experiences:\n output: true\n\n# Defaults\ndefaults:\n - scope:\n path: ""\n type: "publications"\n values:\n layout: "publication"\n - scope:\n path: ""\n type: "experiences"\n values:\n layout: "experience"\n\n# Exclude from processing.\nexclude:\n - .sass-cache/\n - .jekyll-cache/\n - gemfiles/\n - Gemfile\n - G

mcptest-user

Instructions for Setting Up Your Academic Homepage\n\n## Steps to Create Your Personal Homepage\n\n1. Fork the Academic Pages Repository\n - Go to https://github.com/academicpages/academicpages.github.io\n - Click the "Fork" button to create your copy of the repository\n - Rename the repository to "LJT-Homepage" or your preferred name\n\n2. Configure Your Site\n - Edit the _config.yml file with your personal information\n - Update the social media links, email, and description\n - Customize the collections and defaults as needed\n\n3. Update Your Personal Information\n - Modify the index.html or about.md file to include your biography\n - Update the _pages/about.md with your detailed information\n\n4. Add Your Publications\n - Create entries in _publications/ directory for each publication\n - Use the provided BibTeX format or markdown format\n - Include all relevant information (authors, venue, year, links)\n\n5. Add Your Experience\n - Update `_dat

mcptest-user

Publication Files for Academic Pages\n\n## _publications/synlogic.md\n\nmarkdown\n---\ntitle: \"SynLogic: Synthesizing Verifiable Reasoning Data at Scale for Learning Logical Reasoning and Beyond\"\ndate: 2025-01-01\nvenue: 'Arxiv'\npaperurl: ''\ncitation: 'Liu, Junteng, et al. "SynLogic: Synthesizing Verifiable Reasoning Data at Scale for Learning Logical Reasoning and Beyond." Arxiv 2025.'\n---\n\nThis paper presents SynLogic, a method for synthesizing verifiable reasoning data at scale for learning logical reasoning and beyond.\n\n[Code on GitHub](https://github.com/your-repo-link)\n\n\n## _publications/chart-understanding.md\n\n```markdown\n---\ntitle: "On the Perception Bottleneck of VLMs for Chart Understanding"\ndate: 2025-01-01\nvenue: 'Arxiv'\npaperurl: ''\ncitation: 'Liu, Junteng, et al. "On the Perception Bottleneck of VLMs for Chart Understanding." Arxiv 2025.'\n---\n\nThis paper investigates the perception bottleneck of Vision-Language Models (VLMs) for chart underst

mcptest-user

<!DOCTYPE html>
<html lang="en">
<head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
    <title>Junteng Liu - Personal Homepage</title>
    <style>
        body {
            font-family: 'Arial', sans-serif;
            line-height: 1.6;

mcptest-user

# Welcome to Jekyll!
#
# This config file is meant for settings that affect your entire site, values
# which you are expected to set up once and rarely need to edit after that.
# For technical reasons, this file is *NOT* reloaded automatically when you use
# `jekyll serve -l -H localhost`. If you change this file, please restart the 
# server process.

# Basic Site Settings
locale                   : "en-US"

mcptest-user

Junteng Liu's Academic Homepage

This repository contains the files for Junteng Liu's academic homepage, based on the academicpages template.

Files Included

1. _config.yml - Main configuration file with personal information
2. _data/navigation.yml - Navigation menu configuration
3. _pages/about.md - Main about page with personal details
4. _pages/cv.md - CV/Resume page

mcptest-user

---

title: "SynLogic: Synthesizing Verifiable Reasoning Data at Scale for Learning Logical Reasoning and Beyond" collection: publications category: manuscripts permalink: /publication/2025-01-01-synlogic excerpt: 'We introduce SynLogic, a scalable framework for synthesizing verifiable reasoning data to enhance logical reasoning capabilities in large language models.' date: 2025-01-01 venue: 'ArXiv Preprint' paperurl: 'https://arxiv.org/abs/2501.14326' codeurl: 'https://github.com/HKUNLP/SynLogic'

mcptest-user

---

title: "On the Perception Bottleneck of VLMs for Chart Understanding" collection: publications category: manuscripts permalink: /publication/2025-01-01-vision4chart excerpt: 'We investigate the perception bottleneck in vision-language models for chart understanding and propose solutions to improve their performance.' date: 2025-01-01 venue: 'ArXiv Preprint' paperurl: 'https://arxiv.org/abs/2501.08125' codeurl: 'https://github.com/HKUNLP/Vision4Chart'