# Chinchilla Scaling Laws (2022): Hoffmann et al. — 1:1 Param:Data Devrim

> Source: https://sukruyusufkaya.com/learn/llm-muhendisligi/chinchilla-scaling-laws-hoffmann-2022
> Updated: 2026-08-19T19:15:23.620Z
> Category: LLM Mühendisliği
> Module: Modül 12: Scaling Laws — LLM'in Büyüme Kanunları
**TLDR:** Hoffmann et al. 2022 'Training Compute-Optimal LLMs' paper'ı — Kaplan'ı düzeltti. Kaplan undertrained models bias. Chinchilla recipe: N ≈ D (1:1 ratio). 70B Chinchilla model > 280B Gopher (Hoffmann). Llama-3 Chinchilla-aware. Compute-optimal formula yeni, post-Chinchilla overtraining trend.