Ai2 Newsletter
April 2024
Top story - First benchmark for reward models released
Running reward models is hard — but it shouldn't be. RewardBench is here to celebrate success and help us learn more to make it easier. We recently released this first-of-its-kind benchmark to help understand why reward models work. Key takeaways include: more insight into PPO RLHF training, datasets only take us so far, and scaling matters.
$1.48 million grant to help save the whales
The North Atlantic right whale is one of the world’s most endangered large whale species, with approximately only 350 remaining. To turn the tide, our EarthRanger team was awarded a grant of $1.48 million by the National Fish and Wildlife Foundation (NFWF) to bolster ongoing efforts in safeguarding and conserving these incredible animals.
Elaine Zhong is our 2024 Allen AI Outstanding Engineer Scholarship winner!
As part of her scholarship, we're excited to host Elaine Zhong this summer in a paid internship. “This year’s finalists for our 2024 Scholarship were very impressive, much like in years past,” says Dana Young, our Senior Director of People Operations and panel member for the 2024 Allen AI Outstanding Engineer Scholarship. “Elaine stood out out not only on her academic achievements, she stood out with her maturity, determination, and desire to make a real impact in the AI field (with an eye toward linguistics) once she completes her education."
More from us
- OLMo: Open Language Model [Webinar + Q&A]
- How Chain-of-Thought Reasoning Helps Neural Networks Compute
- Is this AI? See if you can spot the technology in your everyday life.