Navreet Kaur
@navreeetkaur
Followers
328
Following
7K
Media
1
Statuses
36
PhD @uw_ischool. Research in NLP, Human-AI Interaction, AI Governance. She/her
Joined March 2016
ICLR authors, want to check if your reviews are likely AI generated? ICLR reviewers, want to check if your paper is likely AI generated? Here are AI detection results for every ICLR paper and review from @pangramlabs! It seems that ~21% of reviews may be AI?
26
95
519
Happily surprised to see OpenAI curating cultural benchmarks, especially focused on India. BUT, cultural knowledge != culturally aligned generations. My work for 2+ years focuses on cultural competence in generative tasks, like creative writing. Sharing some papers in LONG 🧵
Introducing IndQA — a new benchmark that evaluates how well AI systems understand Indian languages and everyday cultural context. https://t.co/MWbRDFQQup
3
7
57
🧐Are values in LLMs aligned with humans? 1️⃣ And if they are — do LLMs stay honest to those values, or sometimes say one thing but act another? 2️⃣ ✨ We explore these questions in two papers presented at #EMNLP2025: 1️⃣ ValueCompass: https://t.co/M4DF2LGg41 (WiNLP Workshop)
1
14
95
I am at #EMNLP2025🇨🇳 to present our main paper *MythTriage: Scalable Detection of Opioid Use Disorder Myths on a Video-Sharing Platform*! Come by to discuss details! 🏦 Location: Hall C ⏲️Time: 11AM-12:30PM 🔗 Paper: https://t.co/Bohi4FxJE3 📁 Repo: https://t.co/2iO6c2MR86
2
2
33
Applying for a #PhD @uw_ischool? Read this 👇 Our student-run application feedback program will be open from October 20th through 1st November 2025. Those from historically underrepresented groups or who have faced barriers in higher ed are highly encouraged to apply. Link 👇
1
4
8
🚨YouTube is a key source of health info, but it’s also rife with dangerous myths on opioid use disorder (OUD), a leading cause of death in the U.S. To understand the scale of such misinformation, our #EMNLP2025 paper introduces MythTriage, a scalable system to detect OUD myth🧵
1
5
16
I will be presenting 👇work at #NAACL2025 tomorrow (May 2) from 12 pm in Ballroom A. Please stop by if curious about inductive biases in transformers, generalization, and applying Bayesian models of cognition for understanding language models.
📢 New Paper! Ever wondered why transformers are able to capture hierarchical structure of human language without incorporating an explicit 🌲 structure in their architecture? In this work we delve deep into understanding hierarchical generalization in transformers. (1/n)
0
6
41
I'll be presenting our work on 'Knowledge Graph-Guided Evaluation of Abstention Techniques' #NAACL 🗓️ Tomorrow @ 9AM 📍Hall 3 Please come say hi + chat about Evals, Human-AI Interaction and AI governance! Paper:
0
2
24
✨Upcoming Talk✨I’ll be speaking at the UW Data Science Seminar tomorrow (April 9) from 4:30pm - 5:20pm (PST) to share ideas about 👫<>🤖“Towards Bidirectional Human-AI Alignment: Empowering Human Interaction in Building Responsible AI across the Lifecycle”. Welcome to hang out
4
10
87
🚀 Are you passionate about #Alignment Research? Exciting news! Join us at the ICLR 2025 Workshop on 👫<>🤖Bidirectional Human-AI Alignment (April 27 or 28, Singapore). We're inviting researchers in AI, HCI, NLP, Speech, Vision, Social Science, and beyond domains to submit their
4
36
159
Remember this study about how LLM generated research ideas were rated to be more novel than expert-written ones? We find a large fraction of such LLM generated proposals (≥ 24%) to be skillfully plagiarized, bypassing inbuilt plagiarism checks and unsuspecting experts. A 🧵
Automating AI research is exciting! But can LLMs actually produce novel, expert-level research ideas? After a year-long study, we obtained the first statistically significant conclusion: LLM-generated ideas are more novel than ideas written by expert human researchers.
32
251
2K
There are a couple of full-time openings for pre-doctoral research associates in my group. One of them is specifically for candidates interested in ensuring that large language (and vision) models are geo-culturally inclusive. Email me if you'd be interested.
4
24
142
I'm at #ACL2024NLP 🇹🇭 presenting our work on how LLMs handle health-related queries with presuppositions ( https://t.co/fnBjQtNWBU) on Tue 12:15 pm (Convention Center A1). Would also love to chat about human-centric evals, conversational systems, AI ethics etc. Come say hi! 😊
New #ACL findings about how LLMs respond to pre-suppositions, particularly for high-stakes health-related queries: https://t.co/86rzJKy82k. Work led by the amazing @navreeetkaur, in collaboration with @monojitchou.
2
3
32
New #ACL findings about how LLMs respond to pre-suppositions, particularly for high-stakes health-related queries: https://t.co/86rzJKy82k. Work led by the amazing @navreeetkaur, in collaboration with @monojitchou.
arxiv.org
As corporations rush to integrate large language models (LLMs) to their search offerings, it is critical that they provide factually accurate information that is robust to any presuppositions that...
2
7
83
In a new AI Snake Oil essay by me and @sayashk, we do a deep dive into AI existential risk probability estimates. We find that these forecasts are just feelings dressed up as numbers, and even the best-run, well funded, time intensive forecasting efforts result in a range of
77
209
800
Folks are surprised that they are giving away gpt-4o for free. They are collecting your speech prompt data silly. Remember if you are not on the team plan, unless you go through a complicated opt-out process, all your data is by default used by openai. Not surprising anymore.
50
171
1K
When will folks realize all the aceing/passing X exams and glimpses of AGI papers are not papers but marketing brochures disguised as “science”? It's like closed LLM companies took a dump on science and the rest of the community is going around with a bucket cleaning it up.
Can AI ace law exams? Last year, OpenAI announced GPT-4 got 90th percentile on the bar. Here we (a) refute 90th percentile claim; (b) replicate/extend recent work on GPT capabilities; (c) discuss implications for law profession. Now open, AI&Law: https://t.co/fX4cwr15gn 1/
4
27
187
At #IIScOpenDay2024, Navreet Kaur (@navreeetkaur) is teasing people about how well they can forecast ChatGPT predictions. And Saksham Rastogi is challenging students to break watermarked language models, to see if they can get away with copying.
Some glimpses from the demonstrations and posters outdoors.
0
5
53
The way we talk about AI matters. “The model understands how to…” implies much more powerful capabilities than “The model is used to…” We present AnthroScore, a measure of how much tech is anthropomorphized, i.e talked about in human-like ways. #EACL2024
https://t.co/VDCW3qB3Gh
5
41
224