(How) Can Transformers Predict Pseudo-Random Numbers?

This paper investigates Transformers' ability to learn pseudo-random sequences from linear congruential generators, revealing their capacity for in-context prediction and generalization to unseen moduli through algorithmic structures. https://arxiv.org/abs//2502.10390 YouTube: https://www.youtube.com/@ArxivPapers TikTok: https://www.tiktok.com/@arxiv_papers Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016 Spotify: https://podcasters.spotify.com/pod/show/arxiv-papers