🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Talk to Juhyun Oh and Eunsu Kim at #eacl about their work on the disparity of LLMs answering q’s and evaluating others’ answers on the same q’s

Great work! I’m glad our benchmarks KoBBQ, KOLD, and CLIcK are used to evaluate HyperCLOVA_X. Making progress in Korean LLM! 🇰🇷🇰🇷🇰🇷 arxiv.org/abs/2307.16778 arxiv.org/abs/2205.11315 arxiv.org/abs/2403.06412

I’m in Bangkok for ACL 2024🇹🇭! We will be sharing KoBBQ in several presentations and poster sessions. If you are interested, please stop by! 12 Aug @ 12:15 (Question Answering I) 14 Aug @ 10:30 (In-Person Poster Session) 16 Aug @ 11:50, 16:00 (C3NLP)

📣Today, I'll present the paper at In-person poster session - at 2pm! Anyone interested in Mechanistic interpretability / Factuality, Let's chat!🙌

Breaking down the Theory-of-Mind task into perception, perspective reasoning, and response generation. Then analyzing the performance of LLMs (pretty bad), led by @ChaniJung99, with Yejin Choi and Hyunwoo Kim, co-authors Jiseon Kim Dongkwan Kim Jiho Jin Yeon Will

FINER is accepted at EMNLP 2024 Main track! Looking forward to meeting everyone at Miami 🌴🌴

🤩Really excited that this work will be presented at #neurips2024 d&b track. The BLEnD dataset took serious collaboration of hard thinking and work, getting human annotations from 16 diverse regional cultures in 13 languages, putting together short-answer and multiple choice QA

I will give a talk at Naver lab (Europe) on Oct 17th, 5 PM (CEST) and 8 AM (PST). This talk is about "how to properly build a metric to evaluate AI-generated text?". I will dive into three main challenges in building a proper evaluation metric and present our proposed

🌟Excited to share our new work! We introduce PROFILE—a framework to uncover LLM misalignments with human preferences and key driving factors. Our study shows LLMs align better as evaluators than generators, and feedback on misalignments improves accuracy. arxiv.org/pdf/2410.06965

Thrilled to share that I'll present my #NeurIPS2024 poster presentation on 11th Dec! "BLEnD: A Benchmark for LLMs on Everyday Knowledge in Diverse Cultures and Languages" arxiv.org/abs/2406.09948 📆 When: Wed 11 Dec 4:30 pm- 7:30 pm 📍 Where: West Ballroom A-D

Can cross-cultural ICL examples help LLM understand underrepresented cultures better? I will present my work "Salad-Bowl-LLM: Multi-Culture LLMs by In-Context Demonstrations from Diverse Cultures" at SoLaR @ NeurIPS2024 (Sat 11am-1pm West 121) #NeurIPS2024 w/ Junho Myung Alice Oh

[1/7] 🚨 New LLM Evaluation Paper Alert! How can we better understand LLMs' abilities? Why not interview them across multiple turns? 🎤 We introduce the LLM-as-an-Interviewer Framework, along with its summarized interview report! 👉 arxiv.org/abs/2412.10424

✨Thrilled to present our paper LLM-C3MOD! 🌍 We tackle the challenge of imbalanced content moderation in cross-cultural settings, and propose a Human-LLM collaborative pipeline to address it. Huge thanks to my co-authors Junyeong Park @ NAACL 2025✈️ Seyoung Song Come say hi at #NAACL #C3NLP!

Jiseon Kim Felipe Vecchietti Alice Oh Mia Cha Thank you for citing our preprint. I'm writing to let you know that it has now been published. Your persona paper is very interesting. I'm currently reading it with students in my lab and we're excited about what we might be able to do next. journals.plos.org/plosone/articl…