(1/5)🚨Can LLMs be more factual without retrieval or finetuning?🤔 -yes✅ 🦙We find factual knowledge often lies in higher layers of LLaMA 💪Contrast high/low layers can amplify factuality & boost TruthfulQA by 12-17% 📝arxiv.org/abs/2309.03883 🧑‍💻github.com/voidism/DoLa #NLProc

📢 NAACL Findings are welcome to submit (non-archival) to the #NAACL2024 trustnlp workshop! Accepted papers will be presented alongside archival submissions. If interested, please fill out the form below ⬇️ ! 📝 Form: forms.gle/TFB9LmkZTRKoii… 🗓️ Deadline: May 1 2024

We are excited to announce that our trustnlp workshop at this year’s NAACL is awarding DEI awards on behalf of our sponsor, Amazon. You can nominate yourself here (forms.gle/vKmMTy3qz8JBj4…). While the award is intended to cover virtual registration fees, you can indicate

We’re thrilled to announce that Dr. Weiyan Shi (Weiyan Shi ) will be a panelist at our upcoming NAACL Student Research Workshop on Tuesday, 06/18! 🌟 The panel will focus on advice for student researchers. 🔗 Learn more about Weiyan Shi: wyshi.github.io

Thrilled to be awarded the prestigious Bloomberg #DataScience Ph.D. Fellowship! 🏆 Grateful for the support and mentorship from Tech At Bloomberg to advance my AI research, especially in LLMs. Heartfelt thanks to Kai-Wei Chang uclanlp & UCLA Computer Science for their tremendous support!

🚨 Contextual Position Encoding (CoPE) 🚨 Context matters! CoPE is a new positional encoding method for transformers that takes into account *context*. - Can "count" distances per head dependent on need, e.g. i-th sentence or paragraph, words, verbs, etc. Not just tokens. -

Want to talk AI fairness? Consider joining as an Indaba mentor and practice digital capacity building, on the ground.

📢 Friends of #Facct2024 , tmrw we are hosting our Queer hackfeminist picnic! 🌅 Welcome women, trans, nonbinary, LGBTQIA* , and allies Down? Be sure to register below 👇🏾 🗓 Date: June 4 📍 Location: Desterro, Rua Santo Amaro 221, Glória 🔗 Register: tinyurl.com/codigohope

Brillantes investigadoras y activistas de f<A+i>r LAC participan de ACM FAccT en Río de Janeiro. ¡Acompañe nuestra agenda! 🌟 Consulta la programación completa de la ACM FAccT Conference: facctconference.org/2024/schedule #FAccT2024

Very excited about this work with Dora Zhao Pooja Chitre, Jerone Andrews, Georgia Panagiotidou, Shawn Walker, Kathleen (Katie) Pine, and Alice Xiang: arxiv.org/abs/2406.06407

📣New! Including citizens throughout the stages of AI development and governance requires a clear vision of what those citizens ideally should *do.* A new primer by Wanheng Hu and Ranjit Singh takes up this imperative. 1/3 datasociety.net/library/enroll…

Student Research Workshop starting today!! Come for the NLP research wisdom starting at 9:30am🧙🏽‍♂️, followed by coffee coffee at 10:30🚀.

Big thanks to our lovely student researchers and panelists for joining today. Your input and energy made it a great event! #NAACL2024

Come to my keynote tomorrow! 9:10 am at the Trustworthy NLP workshop trustnlp. I'll be talking about building reliable NLP tools to empower researchers in text-as-data fields #naacl2024 #naacl24 NAACL HLT 2025

📢Please join us tomorrow for the #NAACL2024 trustnlp workshop, starting at 9AM in Don Alberto 4 ~ 📅 Workshop schedule found below! 🌐But if you can't make it, dont worry - TrustNLP accepted papers are now up on our website, check it out😊trustnlpworkshop.github.io

🤔Are LLMs capable of demonstrating social intelligence in a complex interactive game environment? Check out our recent work: arxiv.org/abs/2406.12203 We introduce ‘InterIntent’ to study social intelligence via intention understanding in the Avalon game. [1/x]

#naacl2024 led to my first time in CDMX and I loved it. Coorganizing NAACL SRW and trustnlp at #naacl2024 was a blast, was stoked for solid reception of my tokenization paper🥹, and incredibly grateful for the mountains just outside the city🌋🏍️. Thank you, Mexico City.

Are you interested in pursuing a Masters or PhD in an AI-related field but unsure where to start? Apply to Black In AI's Emerging Leaders In AI Grad Prep Program. APPLY HERE: bit.ly/3xQG0d2

If AI is the solution to healthcare why did California nurses go on strike against it in April? It's so wild to me how disconnected these two worlds can be. nationalnursesunited.org/press/national…

1/ New paper in Nature shows model collapse as successive model generations models are recursively trained on synthetic data. This is an important result. While many researchers today view synthetic data as AI philosopher’s stone, there is no free lunch. Read more 👇