Zhijing Jin @ACL 2025 Vienna
@ZhijingJin
@zhijingjin.bsky.social☀️Assi. Prof@UofTCompSci. Postdoc@MPI_IS w/ @bschoelkopf. (1) @CausalNLP (2) NLP4SocialGood @NLP4SG (3) Mentor & mentee @ACLMentorship
Congratulations to this wonderful survey paper on cultural diversity in NLP led by @whoSiddheshp @IAugenstein et. al!
Late announcement, but excited to share that our paper was accepted to the Computational Linguistics (CL*) journal! 🎉 Looking forward to presenting at #ACL25 during the oral presentation on Tuesday (Room 1.85, 14:00)
Check out my following sessions at #ACL2025NLP in Vienna:1️⃣Sun 27th: Our ACL Ethics Tutorial 2️⃣Mon 28th 2pm: @ACLMentorship BoF guiding junior researchers in #NLProc 3️⃣Tue 10:30-12: Poster of our "LLM Coding Agent" and 4-5:30pm Poster of our "Multilingual AI Term Dataset" 4️⃣Wed…
Welcome to join our ongoing ACL Ethics tutorial at #ACL2025 “Navigating Ethical Challenges in NLP: Hands-on Strategies for Students and Researchers” in rooms M1-M2! Zoom link available on underline. Hosted by the ACL Ethics Committee to all ACL attendees ❤️ @aclmentorship
Come to meet us and grab chats about mentorship at #ACL2025 in Vienna! @aclmeeting
📢 Join us for the ACL Mentorship Session @aclmeeting #ACL2025NLP #NLProc • Session Link: mentorship.aclweb.org/schedule • Ask Questions: tinyurl.com/y2v2j462 Mentors: • @May_F1_ (@hkust) • @d_aumiller (@cohere) • @vernadankers (@Mila_Quebec) • @ziqiao_ma (@UMichCSE) •…
📢 Join us for the ACL Mentorship Session @aclmeeting #ACL2025NLP #NLProc • Session Link: mentorship.aclweb.org/schedule • Ask Questions: tinyurl.com/y2v2j462 Mentors: • @May_F1_ (@hkust) • @d_aumiller (@cohere) • @vernadankers (@Mila_Quebec) • @ziqiao_ma (@UMichCSE) •…
[Implicit Personalization of #LLMs] How do we answer the question "What colo(u)r is a football?" Answer 1: "Brown🏈 ". Answer 2: "Black and white⚽". We propose a #Causal framework to test if LLMs adjust its answers depending on the cultural background inferred from the question.…
![ZhijingJin's tweet image. [Implicit Personalization of #LLMs] How do we answer the question "What colo(u)r is a football?" Answer 1: "Brown🏈 ". Answer 2: "Black and white⚽". We propose a #Causal framework to test if LLMs adjust its answers depending on the cultural background inferred from the question.…](https://pbs.twimg.com/media/GwJKk5yWAAAEGoP.jpg)
Check out our nice position paper "Causality can systematically address the monsters under the bench(marks)" led by @felixludos @ZhijingJin @bschoelkopf! This work beautifully explains the role of causal structures for #LLM benchmarking 📄arxiv.org/abs/2502.05085 Key insight:…
Check out research opportunities with @lancelotdacosta ! He has very smart insights into math and ML. He’s looking for junior students to join his upcoming projects 👇
Looking for prospective students & collaborators to work towards *An Agent Foundation Model*! 👨💻👩💻 Keywords: Active model discovery, active learning & model based planning. More details: lancelotdacosta.com/research (bottom of page). If you’re interested, get in touch! 💡
1/4 🤝 We can know a lot about a person from their communication style such as the word choice, the organization of arguments and the tone. Similarly, an LLM agent can also exposes their identities from their output. We introduce “interlocutor awareness”—an LLM’s ability to…
The #NLProc community has thought a lot about human Theory of Mind (ToM), what about Multi-Agents? We are excited to introduce "Agent-to-Agent Theory of Mind", where we test whether #LLMs correctly guess each other's identity after a chat, and also show *adapted behavior*…