Training Large Language Models to Reason in a Continuous Latent Space [pdf] | Hacker News

Training Large Language Models to Reason in a Continuous Latent Space [pdf] (arxiv.org)

1 point by thunderbong 2 hours ago | discuss