paper-with-me

홈 › Papers

Compute Optimal Scaling of Skills: Knowledge vs Reasoning

2025-03-13 · Nicholas Roberts, Niladri Chatterji, Sharan Narang, Mike Lewis, Dieuwke Hupkes

Scaling laws are a critical component of the LLM development pipeline, most famously as a way to forecast training decisions such as 'compute-optimally' trading-off parameter count and dataset size, alongside a more recent growing list of other crucial decisions. In this work, we ask whether compute-optimal scaling behaviour can be skill-dependent. In particular, we examine knowledge and reasoning-based skills such as knowledge-based QA and code generation, and we answer this question in the affirmative: $\textbf{scaling laws are skill-dependent}$. Next, to understand whether skill-dependent scaling is an artefact of the pretraining datamix, we conduct an extensive ablation of different datamixes and find that, also when correcting for datamix differences, $\textbf{knowledge and code exhibit fundamental differences in scaling behaviour}$. We conclude with an analysis of how our findings relate to standard compute-optimal scaling using a validation set, and find that $\textbf{a misspecified validation set can impact compute-optimal parameter count by nearly 50%,}$ depending on its skill composition.

📄 PDF Abstract BibTeX arXiv:2503.10061

Code (0)

등록된 구현이 없습니다.

Tasks

Code Generation

Methods 이 논문이 사용한 방법론

SET Dynamic Sparse Training method where weight mask is updated randomly periodically

Similar Papers 제목 키워드 기반

Optimal Sparsity of Mixture-of-Experts Language Models for Reasoning Tasks

2025-08-26 · Taishi Nakamura, Satoki Ishikawa, Masaki Kawamura, Takumi Okamoto 외 arxiv

Empirical scaling laws have driven the evolution of large language models (LLMs), yet their coefficients shift whenever the model architecture or data pipeline changes. Mixture-of-Experts (MoE) models, now standard in st…

Reinforcement Learning

Deliberate Practice: Learning Robot Skills under a Budget

2026-08-13 · Shivam Vats, Sudarshan Harithas, Mete Tuluhan Akbulut, Arvind Raghunathan 외 arxiv

We consider the problem of autonomously learning robot skills under a limited practice budget for sequential tasks. We propose an active skill learning algorithm, \emph{Deliberate Practice (DP)}, that computes a provably…

Are We Scaling the Right Thing? A System Perspective on Test-Time Scaling

2025-09-23 · Youpeng Zhao, Jinpeng LV, Di Wu, Jun Wang 외 arxiv

Test-time scaling (TTS) has recently emerged as a promising direction to exploit the hidden reasoning capabilities of pre-trained large language models (LLMs). However, existing scaling methods narrowly focus on the comp…

LLaDA MoE v2: Scaling Mixture-of-Experts Diffusion Language Models

2026-08-04 · Fengqi Zhu, Shaoxuan Xu, Jingyang Ou, Zebin You 외 arxiv

Diffusion language models (dLLMs) offer an alternative to autoregressive (AR) language modeling, yet the scaling behavior of Mixture-of-Experts (MoE) dLLMs remains poorly understood. We systematically characterize how op…

Towards Thinking-Optimal Scaling of Test-Time Compute for LLM Reasoning

2025-02-25 · Wenkai Yang, Shuming Ma, Yankai Lin, Furu Wei

Recent studies have shown that making a model spend more time thinking through longer Chain of Thoughts (CoTs) enables it to gain significant improvements in complex reasoning tasks. While current researches continue to …

MathMathematical Reasoning