cs.CL · 2026-10-14 · No. 143

Computation and Language, 2026-10-14.

20 new papers in cs.CL. Titles, authors, abstracts. Links to arXiv. Want this in your inbox every morning? Subscribe →

01 — The papers

20 entries
  1. 01

    Tev: Scaling Long-Horizon Agentic Environments with Dense Progress Rewards

    Wenbo Hu, Shraman Pramanick, Krishna Kumar Singh, Nanyun Peng, Kai-Wei Chang, Yong Jae Lee

    cs.CL · cs.AI · cs.CV · cs.LG

    Long-horizon agents have drawn growing attention for their ability to work autonomously -- from hours to days -- on economically valuable tasks, a capability essential to automated discovery and recursive self-improvement (RSI). Yet open-source models still struggle on long-horizon tasks, largely for lack of both training data and a working recipe, and scaling both is hard: long, complex tasks are difficult to curate and verify, and their...

    arxiv.org/abs/2610.18868 · PDF

  2. 02

    DoubtLess: Training-Free Contrastive Decoding for Efficient Reasoning

    Minghao Luo, Liang Chen

    cs.CL · cs.AI · cs.LG

    Reasoning models achieve strong performance through long chains of thought (CoT), but often spend more tokens than needed to solve a problem. Existing ways to shorten the CoT require additional training for each model, add a fixed steering vector to the model's hidden states, or use decoding heuristics such as stopping early or banning words like "wait", which can reduce accuracy. We propose DoubtLess, a training-free decoding method in which...

    arxiv.org/abs/2610.18862 · PDF

  3. 03

    Hybrid Reasoning Agents

    Jiaru Zou, Yunzhe Qi, Sirui Chen, Federico Bianchi, Shang Zhu, Pan Lu, Jingrui He, James Zou

    cs.CL · cs.AI · cs.LG

    How can we build an agent that thinks beyond human language, yet still acts through it in the external environment? Modern LLM agents still reason and interact through text, while latent reasoning offers an efficient internal medium but lacks native tool access and reasoning transparency. We bridge these two directions with Hybrid Reasoning Agents (HybridAgents), which dynamically navigate between latent reasoning and text-based interaction...

    arxiv.org/abs/2610.18861 · PDF

  4. 04

    Back-Translation Still Matters: Leveraging Monolingual Data for Low-Resource Machine Translation with LLMs

    Divyanshu Bathla, Parag Singla, Jitin Singla

    cs.CL · cs.AI

    Pre-trained Large Language Model (LLMs) are effective on many tasks in zero-shot and few-shot settings. For machine translation of low-resource languages, the obvious options are few-shot prompting and fine-tuning on a small parallel set. We ask: what if moderate amounts of low-resource monolingual data are also available? We argue that, rather than continued pretraining, a more effective use of this data is the pre-LLM era technique of...

    arxiv.org/abs/2610.18818 · PDF

  5. 05

    Where Do LLM Ensemble Gains Come From? Decomposing Candidate Value, Diversity, and Aggregation

    Donghwan Kim

    cs.CL · cs.AI

    LLM ensemble gains are often measured against a single baseline, but the value attributed to additional candidates can change with the comparison used. We study an LLM aggregator under four matched conditions: Direct (no candidates), Single-slot (one candidate in a multi-slot interface), Duplicate (one candidate repeated), and Multi-candidate (separately generated candidates). Under either non-diverse control, overall candidate gain (CIV)...

    arxiv.org/abs/2610.18786 · PDF

  6. 06

    From Relation-Specific Heads to Shared Interventions: Contextual Entrainment in Language Models

    Jinghan Zheng, Jiajun Shi

    cs.CL · cs.AI

    Large language models rely on prompt context to perform in-context learning, follow formats, and adapt to local evidence. This ability, however, also raises a robustness concern: models may reuse irrelevant tokens from earlier context when answering a later query. Prior work identifies this behavior as contextual entrainment and finds attention heads whose ablation reduces it, arguing that these heads are largely task- or relation-specific....

    arxiv.org/abs/2610.18674 · PDF

  7. 07

    SMI: Ontology-Grounded Semantic IDs for Hierarchy-Aware Medical Concept Representation

    Ziyang Song, Lia Shen, Yixuan Li, Qincheng Lu, Weihao Li, Haoran Xu, He Zhu, Xinyu Wang, Yue Li

    cs.CL · cs.AI

    Biomedical language models (LMs) provide strong representations for clinical tasks, but they often fail to capture the hierarchical structure among medical concepts. This limitation is especially pronounced for rare and fine-grained clinical codes in long-tailed vocabularies, where hierarchy-agnostic LM representations become less informative as concept frequency decreases. We introduce Semantic Medical Identifier (SMI), a training-free...

    arxiv.org/abs/2610.18672 · PDF

  8. 08

    Known-Voice Cloning and the Limits of Personality Transfer for Character Design

    Johanna Magdalena Kuch, Maud van Lier, Daksitha Senel Withanage Don, Patrick Gebhard, Tanja Schneeberger, Elisabeth André

    cs.CL · cs.AI · cs.HC · cs.RO

    Recent advances in zero-shot voice cloning systems make it possible to reproduce a recognizable voice from a short recording and attach it to arbitrary virtual characters. Prior work suggests that familiar vocal identity can carry over associated personality impressions to new targets such as robots or virtual characters. However, these works have mainly examined familiarity and personalization effects rather than directly testing whether a...

    arxiv.org/abs/2610.18669 · PDF

  9. 09

    Knowledge Graphs as Files: An Agent Harness Where Tool Creation and Tool Use Co-Evolve

    Yuyang Bai, Yu Wang, Yu Zhang

    cs.CL · cs.AI · cs.IR

    Searching text-rich knowledge graphs requires following relations while examining textual evidence. Predefined graph interfaces constrain what evidence agents can access, even when they write code to combine tool calls or process retrieved subgraphs. We propose direct knowledge graph interaction (DKGI), a harness in which language model agents create graph access tools directly over complete node and edge files. The agent writes Python code...

    arxiv.org/abs/2610.18616 · PDF

  10. 10

    Jev-ToM: Evaluating Theory of Mind in Typed Decision Models

    Kawai Chung, Chunkit Chan, Yuxuan Liu, Yauwai Yim, Xi Yang, Wenbin Hu, Huihao Jing, Chiu Yeung Tang, Changxuan Fan,...

    cs.CL · cs.AI

    Dialogue and collaborative assistants repeatedly judge what a partner knows, believes, or intends. Typed decision models offer a potential alternative to generative large language models for bounded Theory of Mind (ToM) selection. We present Jev-ToM, comparing five typed decision models and seven generative baselines on 11,117 English text questions from nine ToM benchmarks across six capability groups. TypeSafe Jev reaches 73.2%...

    arxiv.org/abs/2610.18610 · PDF

  11. 11

    The Many Morals of AI: The cautionary bias of North American news narratives in global perspective

    Andrew Piper, Edie Pearman

    cs.CL · cs.AI

    The stories told about artificial intelligence (AI) play a central role in shaping public understanding, policy debates, and expectations about technological change. News media are especially influential in this process. Using a cross-regional corpus of 12,924 English-language news articles about AI published since 2022, we examine how storytelling about AI varies across media systems in four regions: North America, China, India, and the...

    arxiv.org/abs/2610.18573 · PDF

  12. 12

    The Geometry of Language Model Logits: Generic Structure and Learned Departures

    Swathi Shree Narashiman, Aravind Ramana Venkatramanan, Nambi Seshadri

    cs.CL · cs.AI

    Autoregressive language models produce, at every prediction step, a vocabulary-sized logit vector whose structure determines the conditional next-token distribution. We empirically identify that the ordered vocabulary-head logits follow a remarkably consistent log-linear relation \[ z_{(r)} \approx μ- α\log r. \] across prediction contexts, architectures, and scales, emerging progressively during pretraining. To our knowledge, this is the...

    arxiv.org/abs/2610.18520 · PDF

  13. 13

    Aligning Faithfulness for Theory-Scale Autoformalization

    Yuming Feng, Tadd Mao, Dhruva Arekar, Tianjun Zhong, Jiani Huang, Xujie Si, Ziyang Li

    cs.CL · cs.AI

    Autoformalization (AF) requires more than compilation: formal statements must preserve the meaning of their informal sources, a property known as faithfulness. An emerging frontier is the automated evaluation of this property, yet most existing methods operate at the statement level and produce whole-item verdicts, failing to reveal how individual expressions correspond or which contextual interpretations support the judgment. This omission...

    arxiv.org/abs/2610.18498 · PDF

  14. 14

    Agaram 2.5B: The First Tamil-First Language Model Trained from Scratch with Tool Calling. A Playbook for Under-Represented Languages

    Ramchand Kumaresan, Madhusudhanan G, Kaushik Kezendirane

    cs.CL · cs.AI · cs.LG

    We trained Agaram 2.5B, a 2.5-billion-parameter Tamil-English language model, from scratch on about 50 billion tokens (58% Tamil, the rest English, math, code and parallel text) on one node of eight H100 GPUs in about six days, with a new tokenizer that, unlike our first model's, keeps line breaks. Supervised fine-tuning on 134,428 conversations, whose replies an open-weight teacher model wrote, and three short targeted rounds turned it into...

    arxiv.org/abs/2610.18420 · PDF

  15. 15

    Do Factuality Methods Make LLMs More Factual? Reproducibility and Fair Comparison

    Pratuat Amatya, Vinay Setty, V Venktesh, Avishek Anand

    cs.CL · cs.AI

    Large Language Models (LLMs) are aligned to human preferences, after pre-training and supervised fine-tuning steps. However, standard RL alignment steps such as RLHF do not focus on factuality of LLMs, hence they may generate non-factual information, which could result in unintended consequences. In response, a number of factuality alignment approaches have been proposed in the literature which claim to improve factuality of LLMs on their own...

    arxiv.org/abs/2610.18415 · PDF

  16. 16

    Lightweight but Linguistically Fair? Quantisation Robustness across Bavarian, German, and English

    Jophin John, Michael Hoffmann, Piroska Lendvai, Florent Dufour

    cs.CL · cs.AI

    Post-training quantisation reduces the memory cost of large language models, but it remains unclear whether compression preserves language skills of regional varieties as reliably as their standard-language counterparts. We study this question for Bavarian, Standard German, and English using BAVGROUND, a multiple-choice benchmark, where 206 source questions are aligned across all three languages. Across 14 open-weight 7B--10B instruction...

    arxiv.org/abs/2610.18343 · PDF

  17. 17

    Probing the Geometry of Color Representations in Multilingual Language Models

    Pablo Loyola, Edison Marrese-Taylor, Andrés Hoyos-Idrobo

    cs.CL · cs.LG

    Do multilingual language models learn a universal internal geometry for color, and is it perceptually grounded or primarily linguistic? We decompose color representations into two geometric axes: (P1) alignment to perceptual color space, and (P2) cross-lingual symmetry. Analyzing layerwise embeddings across models and multilingual datasets, we find that perceptual grounding is weak: alignment to CIE-LAB peaks in intermediate layers, then...

    arxiv.org/abs/2610.18331 · PDF

  18. 18

    SimInter: Evidence-Grounded User Simulation with Interview-Based Profile Enrichment for Conversational Recommender Systems

    Erfan Shahabi, Amirhossein Roshandel, Mohammad Sadegh Sadeghi, Azadeh Shakery

    cs.CL · cs.AI · cs.IR

    Conversational Recommender Systems (CRSs) require reliable user simulators for cost-effective evaluation. LLM-based simulators have made realistic simulation possible but typically condition on profiles derived from explicit ratings, demographics, and static sources capturing surface preferences while omitting implicit behavioral characteristics such as decision-making style, aspect-level preferences, and domain expertise. We propose...

    arxiv.org/abs/2610.18308 · PDF

  19. 19

    JevGPT: Thinking Fast and Slow in Autonomous Research

    Yasheng Sun, Yanning Dai, Yimeng Chen, Nanbo Li, Firas Laakom, Yifan Xu, Piotr Piękos, Mateusz Ostaszewski, Wenyi...

    cs.CL · cs.AI · cs.IR

    Autonomous research agents must acquire and reason over evidence within a limited budget. Yet relevant sources can remain uninspected or unused, leaving a gap between retrieval coverage and the evidence that informs research. The challenge is to align fast local judgments with changing evidence needs and feed their findings back into research decisions. We introduce JevGPT, a dual-system architecture for autonomous research that couples fast...

    arxiv.org/abs/2610.18272 · PDF

  20. 20

    TextualRL: Textual Reinforcement Learning for Context Management

    Tianyu Huai, Jie Zhou, Bo Mao, Qin Chen, Liang He, Xipeng Qiu

    cs.CL · cs.AI

    LLM agents can adapt to task feedback by revising the textual context that guides their behavior, without changing their model parameters. However, turning execution rewards into reusable context updates is challenging: individual trajectories offer limited comparative evidence about what to change, while task-specific fixes may overgeneralize when incorporated into a shared context. We introduce TextualRL, a gradient-free reinforcement...

    arxiv.org/abs/2610.18216 · PDF

This edition is part of The Daily Abstract — cs.CL archive. Subscribe to receive these in your inbox each morning, automatically translated to Spanish, with reply-to-PDF: arxivdaily.ignorelist.com.

Colophon Set in Georgia, with system sans for interface chrome and a monospaced stack for code and paper identifiers. Sole accent: amber #D99C5E. Built and served on an always-free VM. The masthead is set 14% letterspaced because newspapers do that and it works.