the way i see it, the last twelve months of AI research can be summed up in just two big breakthroughs:
[i] reasoning ('test-time compute') - new ways to train models that can use more tokens to generate better answers. they mostly rely on RL with verifiable rewards
[ii] memory ('test-time training') - moving towards models that dynamically update weights on-the-fly using gradients. this paper does it by updating a hidden state memory module to 'remember' test data
[i] reasoning ('test-time compute') - new ways to train models that can use more tokens to generate better answers. they mostly rely on RL with verifiable rewards
[ii] memory ('test-time training') - moving towards models that dynamically update weights on-the-fly using gradients. this paper does it by updating a hidden state memory module to 'remember' test data
[i] Learning to reason with LLMs (the o1 blog post)
openai.com/index/learning…
openai.com/index/learning…
[ii] Learning to (Learn at Test Time): RNNs with Expressive Hidden States
arxiv.org/abs/2407.04620
arxiv.org/abs/2407.04620
• • •
Missing some Tweet in this thread? You can try to
force a refresh
