The paper "CompressionBench: How can your coding agents compress LLMs?" investigates whether autonomous coding agents can compress large language models (LLMs) while staying within a specified storage budget. The study frames compression as a terminal task where agents receive a source checkpoint, storage limit, inference environment, and time budget, then generate compressed weights and executable inference code, guided by perplexity feedback and verified by independent checks. Across six agents and six model-size settings, 29 of 36 submissions passed verification, achieving an average 3.4% increase in perplexity over BF16 and 77.33% downstream accuracy, with detailed experiments revealing strengths and failure points in agent-driven compression.
By Xinrui Zhong, Zhifeng Li, Zhiming Wang, Yongqi Liu, Shanglin Yang, Yang Zhou, Zikun Li, Zhuoming Chen, Beidi Chen
The paper introduces ZETA, an architecture that separates routing policy from action execution in scientific vision agents. By resolving scientific intent before invoking specialized visual tools, ZETA treats measurement as a deterministic operation grounded in explicit geometry, enabling independent evaluation of routing correctness and downstream execution. Experiments on a zero‑shot benchmark of 51 scientific‑intent queries demonstrate that explicit routing semantics can dramatically improve routing accuracy, especially for models like Qwen 3.8 where segmentation routing accuracy jumps from 6.2% to 75.0% and detection routing accuracy from 60.0% to 100%.
By Shubhabrata Mukherjee, Jeffrey Donatelli, Daniela Ushizima
The paper studies repeated principal-agent games in a multi-armed bandit setting where the principal offers arm-specific incentives, the agent chooses an arm to maximize its own reward plus incentive, and the principal observes only the chosen arm and its own reward. Because the principal cannot see the agent’s reward, traditional estimation fails, so the authors develop an estimator based on the sequence of incentives and chosen arms, analogous to online inverse optimization. Using this estimator, they design a policy that achieves low (square‑root) regret and validate the approach with simulations for personalized incentive design in long‑term statin treatment.
By Ilgin Dogan, Zuo-Jun Max Shen, Anil Aswani
arXiv:2602.01493v3 Announce Type: replace-cross
Abstract: Solving diverse partial differential equations (PDEs) is fundamental in science and engineering. Large language models (LLMs) have demonstrat...
By Zhuoyuan Wang, Hanjiang Hu, Xiyu Deng, Saviz Mowlavi, Yorie Nakahira
arXiv:2605.18991v3 Announce Type: replace-cross
Abstract: We take the position that agent security must be approached as a systems problem: the AI model powering the agent must be treated as an untru...
By Mihai Christodorescu, Earlence Fernandes, Ashish Hooda, Somesh Jha, Johann Rehberger, Kamalika Chaudhuri, Xiaohan Fu, Khawaja Shams, Guy Amir, Jihye Choi, Sarthak Choudhary, Nils Palumbo, Andrey Labunets, Nishit V. Pandya
arXiv:2609.12097v2 Announce Type: replace-cross
Abstract: In bespoke art commissions, laypeople know what they feel but lack the words to specify it: one participant wanted a laid-off truck driver de...
By Yu-Chao Wang, Yanhong Lu, Yingjie Victor Chen, Tim McGraw
arXiv:2611.05601v1 Announce Type: new
Abstract: A deep RL agent trained on a fixed persona reward plays one way: an Explorer keeps exploring while starving, a Survivor hoards on an empty map. Since e...
By AL Muqshith Mohammed Shifan, Kevin Christopher Chua, Ali Neshati, Loutfouz Zaman, Cristiano Politowski
arXiv:2611.05609v1 Announce Type: new
Abstract: Multi-teacher on-policy distillation (MOPD) extends on-policy distillation (OPD) to consolidate capabilities from multiple teachers. Although prior wor...
By Yichen Wang, Yufa Zhou, Siqi Zhu, Haozheng Luo, Jingzhi Bao, Zhichao Liu, Zijun Wei
arXiv:2611.06084v1 Announce Type: new
Abstract: Hybrid language models interleave a few softmax-attention layers with many linear-attention or state-space layers, so every layer keeps a memory of ear...
By Mingyuan Ma, Yangmin Li, Harris Nover, Zheyu Fu, Po-Han Huang, Alex Settle, Ashwin Nanjappa
arXiv:2611.06704v1 Announce Type: new
Abstract: Training an agent to play Pac-Man with only up, down, left and right actions raises a basic question: when performance is poor, what should be improved...
By Xinglu Wang, Edward Wang, Yong Zhang, Zhenan Fan
arXiv:2611.05580v1 Announce Type: cross
Abstract: We study repeated general games over $T$ rounds between a utility-maximizing optimizer and a learning agent, where budget constraints couple the play...
By Giannis Fikioris, Balasubramanian Sivan, \'Eva Tardos
arXiv:2611.05991v1 Announce Type: cross
Abstract: Multi-agent LLM applications repeat most of their prompt tokens across requests, and serving engines reuse the cached state of shared prefixes. vLLM...
By Muhammad Faizan Anwar, Sina Keshvadi
arXiv:2611.06257v1 Announce Type: cross
Abstract: Open-weight language models that fit on a single consumer GPU are becoming capable on a growing range of tasks, and they matter because they can run...
By Jorge Garc\'ia-Carrasco, Javier Sanchis, \'Alvaro Navarro, Alejandro Mat\'e, Juan Trujillo
arXiv:2509.02861v2 Announce Type: replace
Abstract: The necessary integration of renewable energy sources, combined with the expanding scale of power networks, presents significant challenges in cont...
By Carlo Fabrizio, Gianvito Losapio, Marco Mussi, Alberto Maria Metelli, Marcello Restelli
arXiv:2605.11221v2 Announce Type: replace-cross
Abstract: Predictive models in medicinal chemistry depend on structured assay data locked in primary publications. This bottleneck is acute in targeted...
By Yaochen Rao, Farzaneh Jalalypour, N. M. Anoop Krishnan, Roc\'io Mercado
arXiv:2605.21139v4 Announce Type: replace-cross
Abstract: Current end-to-end autonomous driving models are fundamentally constrained by the behavioral cloning ceiling of imitation learning. While rei...
By Yang Wu, Qiang Meng, Zhaojiang Liu, Youquan Liu, Jian Yang, Jin Xie
arXiv:2611.05720v1 Announce Type: new
Abstract: Search can help LLM agents correct mistakes, but their starting conditions may shape which evidence they find and use. Drawing on cognitive-bias paradi...
By Yanru Xiao, Yingwei Yu
arXiv:2611.06444v1 Announce Type: new
Abstract: Skills let language agents reuse experience across tasks as natural-language documents. Recent work evolves skills from interaction, but existing metho...
By Yanli Wang, Zichen Bai, Tianyue Jiang, Kunming Shao, Junhao Shi, Xueqin Liu, Jinguo Sun, Muye Li, Tingting Bi, Yanlin Wang, Shaoqiang Zheng, Xuefei Diao, Zhuwei Zeng, Nan Duan, Peng Cao
arXiv:2611.06545v1 Announce Type: new
Abstract: Long-running LLM agents compact their context, replacing most history with a summary. We ask whether handoff summaries state permissions or requirement...
By Wenhui Chu
arXiv:2611.06876v1 Announce Type: new
Abstract: LLM agents use context compaction to manage growing interaction histories, but summarizing long histories with large models can introduce substantial d...
By Yulu An, Yifei Feng, Xiangyang Yin