paper-with-me

홈 › Papers

Preserving LLM Capabilities through Calibration Data Curation: From Analysis to Optimization

2025-10-12 · Bowei He, Lihao Yin, Huiling Zhen, Shuqi Liu, Han Wu, Xiaokun Zhang, Mingxuan Yuan, Chen Ma arxiv

Post-training compression has been a widely employed approach to scale down large language model (LLM) and facilitate efficient inference. In various proposed compression methods, including pruning and quantization, calibration data plays a vital role by informing the weight importance and activation dynamic ranges. However, how calibration data impacts the LLM capability after compression is less explored. Few of the existing works, though recognizing the significance of this study, only investigate the language modeling or commonsense reasoning performance degradation from limited angles, like the data sources or sample amounts. More systematic research is still needed to examine the impacts on different LLM capabilities in terms of compositional properties and domain correspondence of calibration data. In this work, we aim at bridging this gap and further analyze underlying influencing mechanisms from the activation pattern perspective. Especially, we explore the calibration data's impacts on high-level complex reasoning capabilities, like math problem solving and code generation. Delving into the underlying mechanism, we find that the representativeness and diversity in activation space more fundamentally determine the quality of calibration data. Finally, we propose a calibration data curation framework based on such observations and analysis, enhancing the performance of existing post-training compression methods on preserving critical LLM capabilities. Our code is provided in \href{https://github.com/BokwaiHo/COLA.git}{Link}.

📄 PDF Abstract BibTeX arXiv:2510.10618

Code (0)

등록된 구현이 없습니다.

Tasks

Code Generation

Similar Papers 제목 키워드 기반

Frequency Matters: Fast Model-Agnostic Data Curation for Pruning and Quantization

2026-03-17 · Francesco Pio Monaco, Elia Cunegatti, Flavio Vella, Giovanni Iacca arxiv

Post-training model compression is essential for enhancing the portability of Large Language Models (LLMs) while preserving their performance. While several compression approaches have been proposed, less emphasis has be…

Model Compression

A Utility-Preserving GAN for Face Obscuration

2019-06-27 · Hanxiang Hao, David Güera, Amy R. Reibman, Edward J. Delp

From TV news to Google StreetView, face obscuration has been used for privacy protection. Due to recent advances in the field of deep learning, obscuration methods such as Gaussian blurring and pixelation are not guarant…

Uncovering Intrinsic Capabilities: A Paradigm for Data Curation in Vision-Language Models

2025-09-27 · Junjie Li, Ziao Wang, Jianghong Ma, Xiaofeng Zhang arxiv

Large vision-language models (VLMs) achieve strong benchmark performance, but controlling their behavior through instruction tuning remains difficult. Reducing the budget of instruction tuning dataset often causes regres…

The Magic Correlations: Understanding Knowledge Transfer from Pretraining to Supervised Fine-Tuning

2026-02-11 · Simin Fan, Dimitris Paparas, Natasha Noy, Binbin Xiong 외 arxiv

Understanding how language model capabilities transfer from pretraining to supervised fine-tuning (SFT) is fundamental to efficient model development and data curation. In this work, we investigate four core questions: R…

Dataset Curation Beyond Accuracy

2021-01-01 · Johan Bjorck, Carla P Gomes

Neural networks are known to be data-hungry, and collecting large labeled datasets is often a crucial step in deep learning deployment. Researchers have studied dataset aspects such as distributional shift and labeling c…

Self-Driving Cars