The prohibitive cost of evaluating Large Language Models (LLMs) necessitates efficient alternatives to full-scale benchmarking. Prevalent approaches address this by identifying a small coreset of items to approximate full-benchmark performance. However, existing methods must estimate a reliable item profile from response patterns across many source models, which becomes statistically unstable when the source pool is small. This dependency is particularly limiting for newly released benchmarks with minimal historical evaluation data. We argue that discrete correctness labels are a lossy view of the model's decision process and fail to capture information encoded in hidden states. To address this, we introduce RepCore, which aligns heterogeneous hidden states into a unified latent space to construct representative coresets. Using these subsets for performance extrapolation, RepCore achieves precise estimation accuracy with as few as ten source models. Experiments on five benchmarks and over 200 models show consistent gains over output-based baselines in ranking correlation and estimation accuracy. Spectral analysis further indicates that the aligned representations contain separable components reflecting broad response tendencies and task-specific reasoning patterns.
Unlocking the Working Memory of Large Language Models for Latent Reasoning
ELLIS Unit Linz and LIT AI Lab, Institute for Machine Learning, Johannes Kepler University Linz, Austria（林茨ELLIS单元和LIT AI实验室，机器学习研究所，约翰·凯撒大学林茨，奥地利）；NXAI GmbH, Linz, Austria（NXAI公司，林茨，奥地利）
Mechanistically Interpreting Compression in Vision-Language Models
Indian Institute of Technology Jodhpur（印度理工学院焦特布尔分校）；AIM Intelligence；National Institute of Technology Agartala（国立阿加尔塔拉理工学院）；Fordham University（福特汉姆大学）；University of Fukui（福井大学）；Independent Researcher（独立研究员）
Chain Of Thought Compression: A Theoretical Analysis
School of Computer and Information Technology, Shanxi University, Taiyuan, Shanxi, China（山西大学计算机与信息学院）；Queen Mary, University of London, UK（伦敦大学女王学院）；School of Informatics, University of Edinburgh, UK（爱丁堡大学信息学院）
TruthX: Alleviating Hallucinations by Editing Large Language Models in Truthful Space
Institute of Computing Technology, Chinese Academy of Sciences（中国科学院计算技术研究所）；University of Chinese Academy of Sciences（中国科学院大学）；Key Laboratory of AI Safety, Chinese Academy of Sciences（中国科学院人工智能安全重点实验室）
Beyond Variance: Knowledge-Aware LLM Compression via Fisher-Aligned Subspace Diagnostics
Department of Computer Science, Iowa State University（计算机科学系，爱荷华州立大学）；Department of Civil, Construction & Environmental Engineering, Iowa State University（土木、建设与环境工程系，爱荷华州立大学）
Implicit Compression Regularization: Concise Reasoning via Internal Shorter Distributions in RL Post-Training
College of Software, Nankai University（南开大学软件学院）；Zhongguancun Academy（中关村学院）；Harbin Institute of Technology（哈尔滨工业大学）；Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；East China Normal University（华东师范大学）；Zhejiang University（浙江大学）；Beijing Institute of Technology（北京理工大学）
ILLC: Iterative Layer-by-Layer Compression for Enhancing Structural Faithfulness in SpArX
Institute for Clarity in Documentation（文献 clarity 研究所）；Inria Paris-Rocquencourt（法国国家信息与自动化研究所巴黎-罗康校区）；Rajiv Gandhi University（拉吉夫·甘地大学）；Tsinghua University（清华大学）；Palmer Research Laboratories（帕尔默研究实验室）；The Kumquat Consortium（金桔联盟）
Gradual Forgetting: Logarithmic Compression for Extending Transformer Context Windows
Department of Computer Science, Indiana University Bloomington（计算机科学系，印第安纳大学布卢明顿分校）；Department of Psychological and Brain Sciences, Indiana University Bloomington（心理学与脑科学系，印第安纳大学布卢明顿分校）
