Chadi Helwe 🇱🇧 🇫🇷

PhD in Artificial Intelligence from @IP_Paris_ Creator of

[@EMNLP عمرو قلج Amr Keleg @Amrkeleg]

🔗 Links:
🗞️ Paper: https://t.co/ycnWRQqajv

📕 VARabiziLex: https://t.co/rzjF375jJ6
🆕 Help us expand the lexicon: https://t.co/5EU4TGhORj

📰 VARabiziCues: https://t.co/YUo84p1Ilg

Co-authors (🙏): @abenabdallah103, @taha_yssne, @Chadi_Helwe, Imane Guellil, and @nedjmaou.

[Mario Nawfal @MarioNawfal]

SAUDI IS BURNING

[Chadi Helwe 🇱🇧 🇫🇷 @Chadi_Helwe]

@uttdkola Djorkaeff

[Chadi Helwe 🇱🇧 🇫🇷 @Chadi_Helwe]

@GunnerPuls Hatem Ben Arfa

[Chadi Helwe 🇱🇧 🇫🇷 @Chadi_Helwe]

@Neduszn8 Di Maria

[François Chollet @fchollet]

The best way to use AI is an interface to information that lets you deepen and improve your own knowledge and mental models. The worst way to use AI is as a crutch to outsource and forsake your own cognition

[Francesco Orabona @bremen79]

Achievement unlocked: I am in a paper with hundreds of authors (together with @Chadi_Helwe and @SultanAlra60920)

[Francesco Orabona @bremen79]

Workshop website: https://t.co/gvDlHh5Yl1

Accepted papers: https://t.co/C8MXNXnjeD

Room: West Ballroom B

[Francesco Orabona @bremen79]

2) We introduce ReviseQA, a new benchmark that evaluates the ability to perform logical reasoning when information changes over multiple conversational turns. Our experiments show that current LLMs often fail to maintain logical consistency when updating beliefs.

[Francesco Orabona @bremen79]

For people at #ICML2025, we have 2 papers in the Workshop on Assessing World Models

1) How are LLMs at playing trading card games (TCG)?Not good once you take care of the contamination! We create a new TGC evaluation task with publicly game engine but hidden card implementation

[@EMNLP عمرو قلج Amr Keleg @Amrkeleg]

TLDR; We are kindly inviting you to participate in our survey—running till the 9th of July—related to Arabizi (sometimes referred to as Arabish or Franco-Arabic): https://t.co/yJ90lFvxMQ

@Amrkeleg Ahmed Amine @taha_yssne Imane Guellil @Chadi_Helwe
@nedjmaou

(1/3) 🧵

[@EMNLP عمرو قلج Amr Keleg @Amrkeleg]

هذه دعوة للمشاركة في استطلاعنا الذي يستمر حتى التاسع من يوليو/تموز والمتعلق بالأرابيزي (يشار إليها أحيانًا باسم أرابيش أو فرانكو أراب): https://t.co/yJ90lFvxMQ

@Amrkeleg أحمد أمين
@taha_yssne إيمان ڨليل
@Chadi_Helwe @nedjmaou

(1/3) 🧵

[Chadi Helwe 🇱🇧 🇫🇷 @Chadi_Helwe]

A huge thank you to Oana Balalau and Davide Ceolin — this wouldn’t have been possible without your collaboration! 🚀

[Chadi Helwe 🇱🇧 🇫🇷 @Chadi_Helwe]

🎉 Excited to share that our paper "Navigating the Political Compass: Evaluating Multilingual LLMs across Languages and Nationalities" has been accepted to the Findings of ACL 2025! 📝

#ACL2025 #LLM #PoliticalBias #MultilingualAI #NLP #Research

[Chadi Helwe 🇱🇧 🇫🇷 @Chadi_Helwe]

Our findings reveal that:
🔹Language plays a key role in shaping the political ideology expressed by a model.
🔹Smaller models tend to have a more stable political ideology than their larger model sizes.

These biases can impact various applications, such as journalism.

[Chadi Helwe 🇱🇧 🇫🇷 @Chadi_Helwe]

@SakanaAILabs @huggingface It would be great to release the merging code :)

[Chadi Helwe 🇱🇧 🇫🇷 @Chadi_Helwe]

Paper link: https://t.co/L6WdiExaYC

[Chadi Helwe 🇱🇧 🇫🇷 @Chadi_Helwe]

🎉 Excited to share the news that my paper titled "MAFALDA: A Benchmark and Comprehensive Study of Fallacy Detection and Classification" has been accepted in the main conference of NAACL 2024! 📝

#NAACL2024 #LLM #AI #NLP #MachineLearning #DeepLearning #NaturalLanguageProcessing

[Chadi Helwe 🇱🇧 🇫🇷 @Chadi_Helwe]

A huge thank you to all my co-authors for their dedication and hard work. This achievement wouldn't have been possible without your contributions! 🚀

[Chadi Helwe 🇱🇧 🇫🇷 @Chadi_Helwe]

We further provide a manual annotation of a part of the dataset together with manual explanations for each annotation. Additionally, we propose a new annotation scheme tailored for subjective NLP tasks and a new evaluation method designed to handle subjectivity.