Study examines how MLPs and Transformers learn noisy k-parity features
A new arXiv paper empirically investigates how MLPs and Transformers differ when trained on datasets containing one or more features encoded by noisy k-parity.
Read the full story at arXiv cs.LG ↗
Timeline · 1 report
- 2026-10-30 04:00 · arXiv cs.LG
Sometimes less is more: Investigating inductive biases of MLPs and Transformers on noisy k-parity features