🔥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

It is awesome to see our work being called (& used) in the message below

Introducing nova.amazon.com, an easier way to experience understanding, content generation, and agentic models from Amazon. #AmazonNova

Excited to present my internship work at Amazon Science at NAACL HLT 2025 tomorrow at 2:00 pm local time. Please come say hi if you are around.

Welcome to the TrustNLP workshop at NAACL25 We are in Rm 215 - San Miguel. Schedule can be found here: trustnlpworkshop.github.io Virtual site access: underline.io/events/492/ses…

#TrustNLP2025, Keynote from ⁦Mor Geva⁩ #NAACL2025

Second Keynote from ⁦Eric Wallace⁩ ⁦trustnlp⁩ #NAACL2025

Keynote 3 by ⁦Niloofar (✈️ ICML)⁩ at ⁦trustnlp⁩ #NAACL2025

We had a great and exciting collaboration with METR on this! Had the opportunity to get great feedback from METR on the new evaluation environment that I developed internally.

🚨 New work: LLMs still struggle at Event Detection due to poor long-context reasoning and inability to follow task constraints, causing precision and recall errors. We introduce DiCoRe — a lightweight 3-stage Divergent-Convergent reasoning framework to fix this.🧵📷 (1/N)

Check out our work with our great Amazon fellow Tanmay Parekh

Happy to share that I have moved to the Bay Area and started as a research scientist at Meta. I will still be working on GenAI safety same as before 🥳

Humans, as social beings, interact, make decisions, and utilize tools. Just as humans exhibit diverse behaviors in various social settings, AI agents can display unpredictable responses when tasked with complex activities or when interacting with one another.