cs.CL · 2026-10-19 · No. 148

Computation and Language, 2026-10-19.

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

01 — The papers

14 entries
  1. 01

    Behavioral Couplings: From Formation to Behavioral Emergence in Language Models

    Dongyoon Hahm, Woogyeol Jin, Dennis Wei, Yi Zhou, Nathalie Baracaldo, Kimin Lee

    cs.CL · cs.AI · cs.LG

    Fine-tuning large language models (LLMs) can induce behaviors beyond those explicitly targeted. For example, training a model to respond more warmly can make it more sycophantic, and training it to write insecure code can cause broad misalignment on prompts unrelated to coding. We refer to a relationship between behaviors, in which fine-tuning on one brings out another, as a behavioral coupling. In this work, we study formation and properties...

    arxiv.org/abs/2610.23847 · PDF

  2. 02

    Beyond the Answer: Student-Conditioned Trace Construction for Reasoning Distillation

    Shaoang Li, Yanhang Shi, Yufei Li, Mingfu Liang, Frank Shyu, Sandeep Pandey, Luke Simon, Xi Liu, Jian Li

    cs.CL · cs.LG

    A reasoning trace can accompany a correct answer without providing effective supervision for a particular student. We study how to construct traces that improve learning beyond answer-only supervision by considering both their intermediate content and their accessibility to the target learner. We introduce ReDiST, a student-conditioned trace-construction and distillation framework for evidence-rich reasoning. Starting from the student's own...

    arxiv.org/abs/2610.23831 · PDF

  3. 03

    Bridging the Language Gap in Scholarly Data II: Enhancing Gender Inference for Chinese Names

    Lisette Espín-Noboa, Mingrong She, Ana Maria Jaramillo, Liuhuaying Yang

    cs.CL · cs.CY · cs.LG · cs.SI · physics.soc-ph

    Perceived gender inference from names is important for inclusive scientometrics research, yet existing methods perform poorly on non-Western scholars. We address this gap for Chinese names, which appear in two forms, Hanzi and Pinyin, its Romanized form (e.g., Zhang Wei, Li Na). In international bibliographic records, names are often available only in Pinyin. However, a single pinyin name can correspond to multiple distinct hanzi characters...

    arxiv.org/abs/2610.23734 · PDF

  4. 04

    GLOBE: Evolving Structured Agentic Memory for E-commerce Product Optimization in Generative Search

    Hyunseo Kim, Suil Choe, Sukkyoung So, Daeyeon Yi

    cs.CL · cs.AI

    Optimizing product listings for generative search requires language agents not only to rewrite effectively but also to continually self-improve by retaining successful strategies across diverse categories. However, existing approaches often treat structured listings as free-form text, leaving effective strategies implicit in prompts, search trajectories, or rewritten outputs rather than accumulating them as a reusable agentic memory. To...

    arxiv.org/abs/2610.23709 · PDF

  5. 05

    Trajectory-Triggered Sparse Collaborative Decoding

    Amit Ranjan Trivedi, Sruthikesh Surineni, Kyong Ho Lee, Mohit Mittal, Steven Li, Rex Hill

    cs.CL · cs.LG

    Dense assistance methods such as speculative decoding repeatedly verify draft tokens with a target model, keeping target computation coupled to generation. We study sparse collaborative decoding, where a draft model generates most tokens and a target model is invoked only for short corrective segments. The key problem is deciding when to call the target. We show that local uncertainty is a weak trigger: entropy, negative log-likelihood, logit...

    arxiv.org/abs/2610.23613 · PDF

  6. 06

    SIT-Eval: Situation-Based Interactive Turn-by-Turn Evaluation of Conversational Agents

    Ravil Mussabayev, Saveliy Zavertyaev, Dmitriy Parpulov, Andrey Savchenko

    cs.CL · cs.AI

    The rapid proliferation of conversational AI agents necessitates evaluation methods that assess performance across diverse dialogue scenarios, utilizing multiple situational criteria, as appropriate behavior depends on context and encompasses various aspects. Existing dialogue benchmarks typically use static natural conversations, which underrepresent rare situations---a critical safety gap, as low-probability events become frequent failures...

    arxiv.org/abs/2610.23612 · PDF

  7. 07

    Truth in Contrast: Calibrated Factual Evidence from Language Model Likelihoods

    Ponhvoan Srey, Jason Lucas, Pittawat Taveekitworachai, Junyoung Lee, Anh Tuan Luu

    cs.CL · cs.AI

    Large Language Models (LLMs) encode substantial factual knowledge. Yet, estimating confidence in factual claims reliably remains challenging. We study likelihood-based factual confidence from the complementary perspectives of discrimination and calibration across multilingual factual statements. Specifically, we instantiate contrastive generative likelihoods as Contrastive Factual Evidence (CFE), which compares the likelihood of the same...

    arxiv.org/abs/2610.23527 · PDF

  8. 08

    HydraLLM: Adaptive Width Slicing for Elastic Text and Vision-Language Models

    Oscar Kemme, Janek Haberer, Ali Hojjat, Olaf Landsiedel

    cs.CL · cs.LG

    LLMs and LVLMs are increasingly deployed across heterogeneous hardware, ranging from data-center GPUs to resource-constrained edge devices. Their fixed-size architectures limit the ability to reuse a single model across deployment settings with different compute, latency, and model-size targets. We introduce HydraLLM, an elastic MQA-based architecture. Its number of active attention heads determines the effective widths of the embeddings,...

    arxiv.org/abs/2610.23455 · PDF

  9. 09

    Sentence-Level Topic Modeling with Lexical and Contextual Evidence

    Federica Granese, Serena Villata, Charles Bouveyron

    cs.CL · cs.LG

    Topic modeling faces a fundamental granularity dilemma: probabilistic topic models assign topics at the word level, whereas transformer-based methods operate at the document level. The latter overlooks that real-world documents typically cover multiple topics, whereas the former relies on individual words, which rarely provide sufficient context to identify them. We introduce SentenceETM, a neural topic model that assigns a topic to each...

    arxiv.org/abs/2610.23388 · PDF

  10. 10

    Language Models Learn to Improve Language Models

    Minhua Lin, Zijun Wang, Zewen Liu, Zhaotian Weng, Bing He, Zhan Shi, Benoit Dumoulin, Xiang Zhang, Suhang Wang, Hanqing Lu

    cs.CL · cs.AI

    Large language models (LLMs) are increasingly used to automate AI research, including language-model (LM) development, offering a potential path toward recursive self-improvement. However, turning experimental feedback into a learned ability to propose effective improvements across models faces two challenges. First, training such an ability with real feedback is expensive: reinforcement learning (RL) requires evaluating tens of thousands of...

    arxiv.org/abs/2610.23344 · PDF

  11. 11

    Think More or Read More? A Token-Budget Theory of Reasoning with Shared Context

    Manh Nguyen, Sunil Gupta, Hung Le

    cs.CL · cs.LG

    Reasoning systems increasingly improve answers by incorporating additional context into the model's input, from multi-agent debate, where agents read their peers' responses, to self-refinement, where a model revisits its own drafts. In these systems, communication consumes context as one agent's output becomes another's input. We study the regime where token reading and generation draw on one fixed token window, so every token spent reading...

    arxiv.org/abs/2610.23246 · PDF

  12. 12

    MGE-PO: Entropy-Aware Credit Assignment for Mathematical Reasoning

    Jihao Huang, Guangjia Chai, Zhirui Chen, Lei Wang, Yao Liu, Liu Qi, Pengjie Ding, Xinyang Liu, Chengxun Jia, Ling...

    cs.CL · cs.AI

    GRPO assigns the same group-relative advantage to equally rewarded responses within a question, leaving their generation uncertainty unused. We introduce MGE-PO, centered on four-quadrant (4Q) advantage shaping that combines verified correctness with rollout generation entropy. Correctness determines the sign of an additional correction, while entropy controls its magnitude: lower-entropy successes receive larger bonuses and higher-entropy...

    arxiv.org/abs/2610.23245 · PDF

  13. 13

    Explaining Chess and Go in Natural Language

    Ori Press, Alexander Golubev

    cs.CL · cs.AI · cs.LG

    Modern chess engines play far beyond human ability, often choosing moves whose underlying ideas are difficult for humans to understand. We propose an automatic metric for evaluating chess explanations, grounded in chess-engine evaluations. Evaluating state-of-the-art language models using this metric, we find that they perform surprisingly poorly, often relying on surface-level board observations rather than robust, transferable chess...

    arxiv.org/abs/2610.23166 · PDF

  14. 14

    Controlling the Diversity-Quality Trade-Off of LLMs by Merging Models Before and After Alignment

    Yasuaki Sumita, Tomoharu Iwata, Toshiyuki Tanaka

    cs.CL · cs.LG

    Alignment is known to change the output characteristics of large language models (LLMs). Base models tend to generate more diverse outputs, whereas aligned models tend to produce higher-quality outputs. The appropriate balance between the two depends on the task. Existing methods either require costly fine-tuning or, like temperature sampling, mainly improve lexical diversity rather than semantic diversity. We introduce a depth-aware...

    arxiv.org/abs/2610.22999 · 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.