Jason Wei
I am an AI researcher based in San Francisco. I currently work at OpenAI. In 2024, I worked on OpenAI o1, a frontier model trained to do chain-of-thought reasoning via reinforcement learning.I gave a video demo of o1-preview for writing puzzles and video game coding, and you can watch a video of OpenAI researchers talking about o1 here.. From 2020 to 2023, I was a research …
Papers - Jason Wei
J. Wei, K. Finn, E. Templeton, T. Wheatley, and S. Vosoughi. NAACL '21 Few-shot text classification with triplet networks, data augmentation, and curriculum learning.
137 emergent abilities of large language models — Jason Wei
2022年11月14日 · Written By Jason Wei Emergent abilities are not present in small models but can be observed in large models. In Emergent abilities of large language models , we defined an emergent ability as an ability that is “not present in small models but is present in large models.”
Thoughts - Jason Wei
Jason Wei. Home Papers Blog Thoughts Open Menu Close Menu. Home Papers Blog Thoughts Thoughts. 2025. Find the right dataset. 2024. Biggest lessons in AI in past five years. Solving hallucinations via self-calibration. Cooking with AI mindset. OpenAI o3. Value of safety research. RL all the time ...
Blog — Jason Wei
2022年11月14日 · For example, in “I’m Jason Wei, a researcher at OpenAI working on large language ___”, it’s not so hard to predict “models”. It’s so easy to predict that token that not much information is lost if I omit it.
Some intuitions about large language models — Jason Wei
2023年11月24日 · Written By Jason Wei An open question these days is why large language models work so well. In this blog post I will discuss six basic intuitions about large language models.
Practicing AI research — Jason Wei
2023年2月7日 · Written By Jason Wei I’m not a particularly experienced researcher (despite my title being “Senior” Research Scientist), but I’ve worked with some talented collaborators and spent a fair amount of time thinking about how to do research, so I thought I might write about how I go about it.
Research I enjoy — Jason Wei
2022年12月29日 · Written By Jason Wei Doing research that is enjoyable is critical to producing outstanding work. Research is a long-term endeavor (done over decades!), involving challenges, failures, and drama.
Successful language model evals — Jason Wei
2024年5月24日 · Written By Jason Wei Everybody uses evaluation benchmarks (“evals”), but I think they deserve more attention than they are currently getting. Evals are incentives for the research community, and breakthroughs are often closely linked …
Common arguments regarding emergent abilities — Jason Wei
2023年5月3日 · Written By Jason Wei This blog post doesn’t represent the positions of my employer (past, present, or future). I’ll review some common arguments that come up when discussing emergent abilities of large language models.