Scalable Complexity Control Facilitates Reasoning Ability of LLMs

• URL: http://arxiv.org/abs/2505.23013v1
• Date: Thu, 29 May 2025 02:42:20 GMT
• Title: Scalable Complexity Control Facilitates Reasoning Ability of LLMs
• Authors: Liangkai Hang, Junjie Yao, Zhiwei Bai, Tianyi Chen, Yang Chen, Rongjie Diao, Hezhou Li, Pengxiao Lin, Zhiwei Wang, Cheng Xu, Zhongwang Zhang, Zhangchen Zhou, Zhiyu Li, Zehao Lin, Kai Chen, Feiyu Xiong, Yaoyu Zhang, Weinan E, Hongkang Yang, Zhi-Qin John Xu,
• Abstract summary: We show that model complexity control can improve the scaling law of large language models consistently over varying model sizes and data sizes.<n>Results indicate that complexity control is a promising direction for the continual advancement of LLMs.
• Score: 41.607173110806265
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The reasoning ability of large language models (LLMs) has been rapidly advancing in recent years, attracting interest in more fundamental approaches that can reliably enhance their generalizability. This work demonstrates that model complexity control, conveniently implementable by adjusting the initialization rate and weight decay coefficient, improves the scaling law of LLMs consistently over varying model sizes and data sizes. This gain is further illustrated by comparing the benchmark performance of 2.4B models pretrained on 1T tokens with different complexity hyperparameters. Instead of fixing the initialization std, we found that a constant initialization rate (the exponent of std) enables the scaling law to descend faster in both model and data sizes. These results indicate that complexity control is a promising direction for the continual advancement of LLMs.

Related papers

• Light-IF: Endowing LLMs with Generalizable Reasoning via Preview and Self-Checking for Complex Instruction Following [10.119219532863767]
  lazy reasoning during the thinking stage is the primary factor contributing to poor instruction adherence.<n>We propose a comprehensive framework designed to enable rigorous reasoning processes involving preview and self-checking.<n>Our Light-IF-32B model surpasses both larger open-source models such as DeepSeek-R1 and closed-source models like Doubao-1.6.
  arXiv  Detail & Related papers  (2025-08-05T07:42:00Z)
• Leveraging Importance Sampling to Detach Alignment Modules from Large Language Models [50.19188692497892]
  Traditional alignment methods often require retraining large pretrained models.<n>We propose a novel textitResidual Alignment Model (textitRAM) that formalizes the alignment process as a type of importance sampling.<n>We develop a resampling algorithm with iterative token-level decoding to address the common first-token latency issue in comparable methods.
  arXiv  Detail & Related papers  (2025-05-26T08:53:02Z)
• Latent Thought Models with Variational Bayes Inference-Time Computation [52.63299874322121]
  Latent Thought Models (LTMs) incorporate explicit latent thought vectors that follow an explicit prior model in latent space.<n>LTMs demonstrate superior sample and parameter efficiency compared to autoregressive models and discrete diffusion models.
  arXiv  Detail & Related papers  (2025-02-03T17:50:34Z)
• A Text-Based Knowledge-Embedded Soft Sensing Modeling Approach for General Industrial Process Tasks Based on Large Language Model [16.842988666530204]
  Data-driven soft sensors (DDSS) have become mainstream methods for predicting key performance indicators in process industries.<n>Development requires complex and costly customized designs tailored to various tasks during the modeling process.<n>We propose a general framework named LLM-TKESS (large language model for text-based knowledge-embedded soft sensing) for enhanced soft sensing modeling.
  arXiv  Detail & Related papers  (2025-01-09T08:59:14Z)
• Byte Latent Transformer: Patches Scale Better Than Tokens [101.10994909832063]
  Byte Latent Transformer (BLT) encodes bytes into dynamically sized patches, which serve as the primary units of computation.<n>For fixed inference costs, BLT shows significantly better scaling than tokenization-based models, by simultaneously growing both patch and model size.
  arXiv  Detail & Related papers  (2024-12-13T05:33:32Z)
• Can a Large Language Model Learn Matrix Functions In Context? [3.7478782183628634]
  Large Language Models (LLMs) have demonstrated the ability to solve complex tasks through In-Context Learning (ICL) This paper explores the capacity of LLMs to solve non-linear numerical computations, with specific emphasis on functions of the Singular Value Decomposition.
  arXiv  Detail & Related papers  (2024-11-24T00:33:43Z)
• Unlock the Correlation between Supervised Fine-Tuning and Reinforcement Learning in Training Code Large Language Models [12.656574142412484]
  We make an attempt to understand the correlation between supervised fine-tuning and reinforcement learning.<n>We find that both atomic and synthetic functions are indispensable for SFT's generalization.
  arXiv  Detail & Related papers  (2024-06-14T03:39:01Z)
• Curated LLM: Synergy of LLMs and Data Curation for tabular augmentation in low-data regimes [57.62036621319563]
  We introduce CLLM, which leverages the prior knowledge of Large Language Models (LLMs) for data augmentation in the low-data regime. We demonstrate the superior performance of CLLM in the low-data regime compared to conventional generators.
  arXiv  Detail & Related papers  (2023-12-19T12:34:46Z)
• To Repeat or Not To Repeat: Insights from Scaling LLM under Token-Crisis [50.31589712761807]
  Large language models (LLMs) are notoriously token-hungry during pre-training, and high-quality text data on the web is approaching its scaling limit for LLMs. We investigate the consequences of repeating pre-training data, revealing that the model is susceptible to overfitting. Second, we examine the key factors contributing to multi-epoch degradation, finding that significant factors include dataset size, model parameters, and training objectives.
  arXiv  Detail & Related papers  (2023-05-22T17:02:15Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.