pith. sign in

Transformers learn to implement multi-step gradient descent with chain of thought

3 Pith papers cite this work. Polarity classification is still indexing.

3 Pith papers citing it

years

2026 2 2025 1

clear filters

representative citing papers

The Power of Power Law: Asymmetry Enables Compositional Reasoning

cs.AI · 2026-04-24 · unverdicted · novelty 6.0

Power-law data sampling creates beneficial asymmetry in the loss landscape that lets models acquire high-frequency skill compositions first, enabling more efficient learning of rare long-tail skills than uniform distributions.

citing papers explorer

Showing 2 of 2 citing papers after filters.

  • Transformers Provably Learn to Internalize Chain-of-Thought cs.LG · 2026-05-27 · unverdicted · none · ref 16

    L-layer transformers under Log-ICoT curriculum provably learn k-parity with poly(n) samples and log k stages, matching explicit CoT efficiency without inference overhead.

  • The Power of Power Law: Asymmetry Enables Compositional Reasoning cs.AI · 2026-04-24 · unverdicted · none · ref 23

    Power-law data sampling creates beneficial asymmetry in the loss landscape that lets models acquire high-frequency skill compositions first, enabling more efficient learning of rare long-tail skills than uniform distributions.