cs.AI · cs.CL · cs.LG · cs.CV · cs.CR · cs.SE · stat.ML · q-fin.TR · q-fin.ST · econ.GN · q-bio.QM · physics.soc-ph · math.OC · cs.RO · cs.HC
Qingming Li, Ruiyan Sun, Satoshi Nakamura · 2026-10-20
Picture-description dementia classification uses a shared visual scene, but a transcript-level prediction alone does not identify the scene evidence underlying its inputs. NarraSpace links words and speech times to named picture elements and retrieves utterances for predefined sc…
Noor Abo Mokh, Teresa Lynn, Nizar Habash, Ahmed Elshabrawy, Sara Alhajeri and others · 2026-10-20
Mathematical reasoning remains underexplored for low-resource languages, including Dialectal Arabic. To bridge this gap, we introduce the DAR-MGSM (Dialectal Arabic Multilingual Grade School Math) Benchmark, the first manually curated benchmark for evaluating mathematical reasoni…
Shu Zhao, Tan Yu · 2026-10-20
Agentic on-policy distillation (OPD) trains a student agent on its own trajectories while a teacher scores every token. A natural explanation of where it goes wrong is compounding error: the student's mistakes carry its trajectories beyond the region where the teacher is reliable…
Tianyu Fu, Ruoxi Wang, Zihan Min, Zhongqi Ren, Zhihao Li and others · 2026-10-20
Large Language Model (LLM) agents can handle increasingly complex tasks, navigating long trajectories of observations and actions. As these trajectories grow, key-value (KV) cache I/O and storage become the dominant cost. Trajectories can also exceed the LLM's maximum context len…
Arnav Chavan, Chanhyuk Lee, Udbhav Bamba, Jaehoon Yoo, Seunghoon Hong and others · 2026-10-20
Continuous flow language models offer a path to parallel generation with tools for control and acceleration unavailable to discrete diffusion, yet effective techniques for building them at the billion-parameter scale remain less established. We present Lafa (Large flow language m…
Rodrigo Casado Noguerales, Dimitri von Rütte, Bernhard Schölkopf, Thomas Hofmann, Aran Raoufi · 2026-10-20
Discrete diffusion models iteratively denoise a noisy sequence of categorical data such as text, enabling any-order generation with self-correction and in-filling capabilities. This makes them promising candidates for reasoning, planning, and constrained-generation tasks like cod…
Ning Zhang, Shihao Wang, Jinrui Zhang, Chaodong Xiao, Lei Zhang · 2026-10-20
For efficient deployment of large language models (LLMs), methods of ultra-low-bit backbones reduce weight storage, memory traffic, and linear-layer execution cost, while multi-token prediction (MTP) methods reduce sequential decoding steps by predicting multiple future tokens in…
Aur Shalev Merin · 2026-10-20
Probing studies of latent reasoning often read a model's state where the answer begins and take a failed probe as evidence that a value is absent. We test this with a digit copied from the question, in the looped model Ouro-1.4B and the standard transformer Qwen2.5-1.5B. Comparin…
Ruoyu Zhang, Zhaocen Liu, Weiyi Tian, Melissa Feng, Shilong Liu · 2026-10-20
Long-term assistants must follow unfinished tasks across days as information and circumstances change. Finding information or starting an action, however, does not ensure that the task gets completed. We introduce OptionWorld, a benchmark that traces information acquisition, judg…
Cheng Liang, Tengfei Zhang, Wenjie DU, Jianyu Wu, Changlian Ma and others · 2026-10-20
Looped Transformers increase effective depth by repeatedly applying shared layers, offering a way to scale computation without increasing parameter count. However, full-stack looping places context construction and iterative state refinement within the same recurrent computation,…
Ziyong Wang, Yupei Lin, Runxuan Huang, Liang Lin, Dongyu Zhang and others · 2026-10-20
General-purpose multimodal models have shown promise as direct robot action policies, yet whether they grasp the physics underlying real-world interactions and can apply that understanding to guide action remains underexplored. We investigate this question using contact-rich mani…
Zhenwei Tang, Di Bai, Jintao Liu, Xidong Wu, Qianfeng Wen and others · 2026-10-20
Agents are starting to sell to other agents. Seller agents write the listings of tools, services and products, and buyer agents read these listings when they choose what to use. We show that seller agents told only to be selected more often practice generative engine optimization…
Nikita Ganziuk, Innokentiy S. Humonen, Ilya Makarov, Pavel Lebedev · 2026-10-20
We present a tool for the comparative study of mythological and religious texts. Traditional manual analysis is highly labor-intensive and difficult to scale. To address this, we use neural language models to embed texts into a shared semantic space, so that users can find semant…
Mohammad Reza Modarres, Armin Tourajmehr, Artin Tavasoli, Yadollah Yaghoobzadeh, Mohammad Taher Pilehvar · 2026-10-20
Large language models leave model-specific traces in their generated text, but it remains unclear how finely these fingerprints can distinguish related models and which signals remain informative as obvious cues are reduced. We study this through fingerprint resolution, consideri…
Christophe Khalil, Aya Mourad · 2026-10-20
We present our solution to the AKBC LM-KBC 2026 Shared Task, where we achieved a score of 0.6836 on the official leaderboard, more than double the baseline of 0.2964. This task requires generating all possible objects for subject-relation pairs using a language model with up to 3…
Mohammad Kachuee, Phu Mon Htut, Sheng Li, Nikolaos Pappas, Miguel Ballesteros Martinez and others · 2026-10-20
Agentic AI extends large language models (LLMs) from generating responses to planning and executing complex workflows, substantially expanding their utility while exposing applications to unsafe actions, information leakage, and adversarial manipulation. Multi-agent systems furth…
Jinghui Zhang, Lang Gao, Zongfang Liu, Ruihong Zeng, Zirui Song and others · 2026-10-20
Precisely reproducing a target span verbatim within a long context is a fundamental capability test for LLMs: even when the target string is present in context, models often still drift within the span or fail to stop correctly, revealing limitations in source grounding and bound…
Youmna Farag, Svetlana Stoyanchev, Simon Keizer, Mohan Li, Rama Doddipatla · 2026-10-20
Reasoning about safety is essential for embodied agents powered by large language models (LLMs). We introduce the Contextual Action and Plan Safety benchmark (CAPSafe), designed to evaluate safety reasoning at two levels: (1) an action-level dataset that tests whether a single ac…
Gaotang Li, Bhavana Dalvi Mishra, Yifei Ming, Yanfei Chen, Zhongying CuiZhu and others · 2026-10-20
Agent harnesses are executable code that specifies the prompts, tools, and control flow around a language model. A single global harness can struggle with heterogeneous tasks that require different execution strategies. In this work, we introduce Object-Oriented Harness (OO-Harne…
Atharva Thakur, Abulhair Saparov, Bruno Ribeiro · 2026-10-20
Reliable retrieval-augmented generation (RAG) requires recognizing when neither retrieved documents nor parametric knowledge supports an answer. We introduce DeclineCraft, a dynamic abstention benchmark whose unanswerable questions target Minecraft feature additions after a corpu…
Rohan James, Ioannis Christofilogiannis, Yuuv Jauhari, Shreya Shah, Gauri Pethkar · 2026-10-20
Emoji are widely treated as compact affective tokens whose meaning is summarised by a single sentiment score. We argue that this picture is incomplete. From a corpus of 16{,}575 English tweets and a per-emoji metric we call \emph{stability delta} (within-class minus between-class…
Masahiro Kaneko, Hiromi Arai, Timothy Baldwin · 2026-10-20
Large language model (LLM) agents increasingly use persistent memory to personalize responses, and personal attributes in this memory pose a privacy risk if exposed. We ask how consulting memory during dialogue changes what users go on to say, and thus what the memory comes to re…
Amit Lisha, Reut Tsarfaty · 2026-10-20
Natural-language instructions often rely on high-level abstractions, such as asking an agent to ``stack blocks in a pyramid.'' Interpreting abstract instructions requires models to both identify the abstraction intended by the speaker, and unroll it into the many concrete actions…
Tianyi Lorena Yan, John Hewitt · 2026-10-20
Monitoring the chain-of-thought (CoT) of language models (LMs) is a key strategy for detecting and preventing misbehaviors for AI safety. As such, it is critical to identify reinforcement learning (RL) practices that lead to unmonitorable chain-of-thought. One such practice may b…
Ahsan Bilal, Muhammad Ahmed Mohsin, Muhammad Umer, Ayesha Mohsin, Sarmad Bilal and others · 2026-10-20
Test-time scaling allocates additional inference compute to refining existing solutions or generating new ones. This raises a fundamental question: Should I keep improving this solution, or stop working on it and try a completely new solution? Existing methods make this decision …
Tapan Parikh · 2026-10-20
A model behind an API can be identified from its one-word answers to everyday questions such as "Name a sound" or "Name a poet". A model's profile records how often it gives each answer to each question. We attribute new answers to the model most likely to have given them. On a p…
Tapan Parikh · 2026-10-20
Asked to name a running shoe brand, language models say Nike. Asked to recommend one, they say Brooks. This paper asks where models learn that taste. Across 102 models and 42 brand categories, the brand a model picks when recommending differs from the brand it names in 26. The fi…
Pratik Worah, Randall Bateman · 2026-10-20
Progressive neuronal loss is the hallmark of neurodegenerative disease, yet how loss of neurons shapes the pattern of cognitive decline remains poorly understood. Here we systematically ablate neurons (computational units) in language models spanning recurrent long short-term mem…
Tzu-Quan Lin, Chih-Kai Yang, Yi-Cheng Lin, Hao Tang, Hung-yi Lee · 2026-10-20
While the language model backbone is pretrained on text alone, large audio-language models (LALMs) can, to a great extent, recognize acoustic attributes, such as emotion, speaker gender, and spoken language. However, it is unclear how the audio representation is integrated, if at…
Jiapu Wang · 2026-10-20
Large Language Models (LLMs) often struggle with temporally grounded reasoning because their static parametric memory is misaligned with the dynamic evolution of real-world knowledge. Although RAG and Parameter-Efficient Fine-Tuning (PEFT) provide practical ways to enhance LLMs, …
Mich-Seth Owusu · 2026-10-20
Large Language Models (LLMs) depend fundamentally on the diversity, depth, and factual rigor of their pre-training corpora. However, contemporary foundational pre-training datasets (such as The Pile, S2ORC, and Dolma) exhibit a severe structural omission: African scientific, medi…
R. Vestrelli, A. Fronzetti Colladon · 2026-10-20
Changes in occupational task content may have different economic implications depending on whether the tasks gained or lost are occupation-specific or already widespread across the economy. Existing measures of occupational task change largely abstract from this distinction, trea…
Stephanie Fong, Yiwen Jiang, Qingyang Xu, Zimu Wang, Yaling Shen and others · 2026-10-20
Large Language Models (LLMs) are increasingly used for interpersonal advice, potentially shaping how users interpret ambiguous social encounters. Existing Theory of Mind evaluations typically rely on predefined answers, while uncertainty research overlooks LLMs' preferred explana…
Yousra Fettach, Tijl De Bie · 2026-10-20
Persuasive writing exploits cognitive biases, recruiting a reader's mental shortcuts to shape belief, often without the reader noticing. We call the linguistic features exploiting such mental shortcuts cognitive bias triggers. As large language models (LLMs) increasingly mediate …
Devansh Srivastav, David Pape, Samira Abedini, Lea Schönherr · 2026-10-20
LLMs increasingly operate over mixed-sensitivity contexts, where they must reason over information ranging from public to confidential while disclosing only what the user is authorised to receive. Existing approaches enforce access control as a binary decision: either fully answe…
Haonan Huang, Yuwei Lu, Chengkun Cai · 2026-10-20
Multilingual evaluations often score a model's opening yes or no, a token that also carries model and language answering conventions. We ask when different ways of eliciting and reading an answer recover the same judgment, and when they measure an answer habit. We collect 1.1 mil…
Runyang You, Yongqi Li, Tiezheng Yu, Wenjie Li · 2026-10-20
On-policy distillation (OPD) consolidates expert capabilities into a single student by matching teacher and student final-layer distributions on student-generated rollouts. We investigate whether depthwise prediction transitions, the changes in predictions between layers, can pro…
Masahiro Kaneko · 2026-10-20
To defend against backdoor attacks on neural networks, the defender must identify the verifier function that determines whether an input is a trigger, rather than rejecting individual triggers in isolation, since an attacker can generate new triggers satisfying the same verifier …
Michael Dalvean · 2026-10-20
We use Letter Positional Probabilities (LPPs) to examine the orthographic differences between simple and complex French words. We draw from a sample of 38,000 French words scored by reaction time on a lexical decision task. We take the 5000 highest reaction time words and the 500…
Taiheng Pan · 2026-10-20
Practitioners use reinforcement learning methods for language models through training frameworks, so a method that a framework includes is one configuration change away for its users. We measure how often this happens. We trace seven open-source frameworks for RL post-training th…
GET https://store.agentexchange.work/gw/papers/latest?cat=cs.CL&limit=25 (0.003 USD per call). Catalogue