1. X
  2. Haorui He
Log inSign up
Haorui He
113 posts
user avatar
Haorui He
@HeHarry_11
CS Ph.D. student at HKU @HKUniversity; LLM/Social Computing/NLP; Fan @BVB
Hong Kong
harryhe11.github.io
Joined April 2021
454
Following
86
Followers
RepliesRepliesMediaMedia

Log in or sign up for X

See what’s happening and join the conversation

Continue with phone
or
Log in with username or email
Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
  • user avatar
    Haorui He
    @HeHarry_11
    Jul 28
    1/6 🎉 Our work “Novel Claim or Déjà Vu? Rethinking ‘Contamination-Free’ Dynamic Evaluation for Multimodal Automated Fact-Checking” accepted by ACM MM 2026 #ACMMM is now officially open-sourced! 💻 Code: github.com/TrustworthyCom… 📑 Paper: arxiv.org/abs/2607.23514
    GitHub - TrustworthyComp/Rethink-MAFC-Eval: Official code for ACM MM 2026 paper: Novel Claim or...
    From github.com
    224
  • user avatar
    Haorui He
    @HeHarry_11
    Oct 2, 2025
    The real OpenAI!
    user avatar
    Zichen Liu
    @zzlccc
    Oct 2, 2025
    much more convinced after getting my own results: LoRA with rank=1 learns (and generalizes) as well as full-tuning while saving 43% vRAM usage! allows me to RL bigger models with limited resources😆 script: github.com/sail-sg/oat/bl…
    363
  • user avatar
    Haorui He
    @HeHarry_11
    Aug 22, 2025
    Compressed Emilia-Yodas from 1.7TB to 41 GB 👍👍👍
    user avatar
    Jiameng
    @JiamengJiameng
    Aug 14, 2025
    We're open-sourcing our Neural Audio Codec: huggingface.co/neuphonic/neuc… NeuCodec is a 24kHz single codebook codec running at 0.8kbps. It's what we use in production and it's awesome
    280
  • user avatar
    Haorui He
    @HeHarry_11
    Aug 14, 2025
    🚨 Excited to introduce Fact2Fiction: A novel poisoning attack targeting agentic fact-checking systems! 😈 It exploits agentic systems to craft malicious evidence and boosts attack success by up to 21.2% over prior SOTA attacks! Details: arxiv.org/abs/2508.06059 #FactChecking #LLM
    arxiv.org
    Fact2Fiction: Targeted Poisoning Attack to Agentic Fact-checking System
    State-of-the-art (SOTA) fact-checking systems combat misinformation by employing autonomous LLM-based agents to decompose complex claims into smaller sub-claims, verify each sub-claim...
    179
  • user avatar
    Haorui He
    @HeHarry_11
    Jul 19, 2025
    Wow!!! Emilia with emotion labels from Laion! Winning of OPEN-SOURCE! 🎉@ChristophSchuh6 @laion_ai @realmrfakename Congrats 🎊
    user avatar
    Christoph Schuhmann
    @ChristophSchuh6
    Jul 18, 2025
    Yay! 🥳 We are done with Emilia! 215k hours of emotion annotated, multilingual speech! huggingface.co/datasets/laion… huggingface.co/datasets/laion… ... huggingface.co/datasets/laion…
    218