cs.CL · 2026-06-16 · No. 25

Computation and Language, 2026-06-16.

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

01 — The papers

19 entries
  1. 01

    KVEraser: Learning to Steer KV Cache for Efficient Localized Context Erasing

    Mufei Li, Shikun Liu, Dongqi Fu, Haoyu Wang, Yinglong Xia, Hong Li, Hong Yan, Pan Li

    cs.CL · cs.LG

    Post-hoc context erasing over the KV cache is challenging because a local edit has a global consequence: once a span has been processed, its influence propagates into the cached states of all subsequent tokens. This issue arises naturally in long-context LLM applications, where stale retrieved facts, incorrect tool observations, retracted user preferences, or harmful prompt injections may be identified only after prefill. Exact erasing must...

    arxiv.org/abs/2606.17034 · PDF

  2. 02

    TokenPilot: Cache-Efficient Context Management for LLM Agents

    Buqiang Xu, Zirui Xue, Dianmou Chen, Chenyang Fu, Chiyu Wu, Caiying Huang, Chen Jiang, Jizhan Fang, Xinle Deng,...

    cs.CL · cs.AI · cs.LG · cs.MA

    As LLM agents are deployed in long-horizon sessions, context accumulation drives up inference costs. Existing approaches utilize text pruning or dynamic memory eviction to minimize token footprints; however, their unconstrained sequence mutations alter layouts, introducing prefix mismatches and cache invalidation. This reveals a critical trade-off between text sparsity and prompt cache continuity. To address this, we present TokenPilot, a...

    arxiv.org/abs/2606.17016 · PDF

  3. 03

    Exploring Extrinsic and Intrinsic Properties for Effective Reasoning with Code Interpreter

    Patomporn Payoungkhamdee, Napat Laosaengpha, Jenta Wonglertsakul, Pittawat Taveekitworachai, Pume Tuchinda,...

    cs.CL · cs.LG

    Reasoning with a Code Interpreter (CI) has emerged as an effective paradigm for enhancing the reasoning capabilities of large language models (LLMs) through executable computation and iterative verification. Despite its growing adoption, the behavioral properties underlying effective code reasoning remain largely underexplored. In this work, we investigate code reasoning from two distinct perspectives inspired by prior studies of natural...

    arxiv.org/abs/2606.16934 · PDF

  4. 04

    IMPACTeen: Intentions, Manipulation, Persuasion, Annotations, and Consequences in Teen Communication Dataset

    Aleksander Szczęsny, Wiktoria Mieleszczenko-Kowszewicz, Maciej Markiewicz, Beata Bajcar, Tomasz Adamczyk, Jolanta...

    cs.CL · cs.AI

    IMPACTeen is a dataset of textual social influence scenarios spanning interpersonal, media-based, and digital settings in an adolescent context. It contains 1,021 texts, 5,100 individual annotation records, and gold labels for social influence techniques, with each text annotated from five distinct perspectives: teenagers, parents, psychologists, communication experts, and teachers. The resource was constructed through constrained LLM...

    arxiv.org/abs/2606.16910 · PDF

  5. 05

    Compositional Reasoning Depth Predicts Clinical AI Failure: Empirical Evidence Consistent with Transformer Compositionality Limits in Electronic Health Record Question Answering

    Sanjay Basu

    cs.CL · cs.AI

    Aggregate accuracy benchmarks conceal a systematic structure in how large language models fail at electronic health record (EHR) question answering: questions requiring more inferential steps produce disproportionately more errors. Motivated by theoretical results on transformer compositionality limits, we introduce a pre-specified hop-count taxonomy -- the number of distinct reasoning steps required to answer a clinical question from an EHR...

    arxiv.org/abs/2606.16890 · PDF

  6. 06

    Follow the Latent Roadmap: Navigating Revocable Decoding for Diffusion LLMs with Anchor Tokens

    Yizhen Yao, Qinglin Zhu, Runcong Zhao, Xiangxiang Dai, Yanzheng Xiang, Yulan He, Lin Gui

    cs.CL · cs.AI

    Diffusion Large Language Models (dLLMs) offer a promising avenue for parallel generation but face a trade-off between decoding speed and quality. While revocable decoding strategies attempt to mitigate errors by verifying and remasking tokens, they typically operate within a mixed-quality context. This leads to two critical failures: \textit{Error Propagation}, where new tokens absorb toxic information from erroneous context, and...

    arxiv.org/abs/2606.16847 · PDF

  7. 07

    Robust Dual-Signal Fusion: Hybrid Neuro-Symbolic Gating with Compressed Chain-of-Thought Refinement for Irony Detection in Social Media Texts

    Ankit Bhattacharjee, Krityapriya Bhaumik

    cs.CL · cs.AI

    Large Language Models (LLMs) natively default to literal semantic interpretations, making zero-shot irony detection a persistent challenge. We introduce the Robust Dual-Signal (RDS) Fusion framework, a hybrid neuro-symbolic architecture that compresses Chain-of-Thought (CoT) reasoning trajectories without Supervised Fine-Tuning (SFT). Evaluated on a strictly held-out TweetEval test set (N=734), RDS achieves 78.1% accuracy and a Macro F1 of...

    arxiv.org/abs/2606.16845 · PDF

  8. 08

    Tying the Loop -- Tied Expert Layers in Mixture-of-Experts Language Models

    Martin Jaggi

    cs.CL · cs.AI · cs.LG

    Mixture-of-Experts (MoE) architectures efficiently scale Large Language Models (LLMs) by activating only a small fraction of their experts per token, yet the full parameter count - dominated by the expert parameters - must be held in training and inference memory. To address this, we introduce Expert Tying, an architectural modification that shares expert parameters across consecutive transformer layers while preserving independent,...

    arxiv.org/abs/2606.16825 · PDF

  9. 09

    P3B3: A Multi-Turn Conversational Benchmark for Measuring European and Brazilian Portuguese Variety Bias in LLMs

    Rafael Ferreira, Inês Vieira, Inês Calvo, James Furtado, Iago Paulo, Diogo Tavares, Diogo Glória-Silva, David...

    cs.CL · cs.AI · cs.LG

    As Large Language Models (LLMs) become embedded in everyday communication, capturing regional linguistic variation is essential for reliable and equitable language use. In Portuguese, European (pt-PT) and Brazilian (pt-BR) varieties remain unevenly represented, with pt-BR dominating in data quantity, while LLM preference for Portuguese variants remains underexplored. To address this gap, we introduce P3B3, an expert-curated language variety...

    arxiv.org/abs/2606.16753 · PDF

  10. 10

    Sycophancy as Material Failure under Pushback Loading: A Multi-Axis Characterization Across Three Loading Cases and up to Seventeen Material Charges

    Ferdinand M. Schessl

    cs.CL · cond-mat.mtrl-sci · cs.AI

    Sycophancy in LLMs is documented across 70+ papers, but expert agreement on construct boundaries remains low (ICC=.184; Ye et al., 2026). The construct fragments because behavioral classification depends on which surface form is privileged. We adopt a materials-science framing: conversation as test specimen under load, LLM-model as material charge, pushback as progressive load, stance-flip as material failure. We characterize this failure...

    arxiv.org/abs/2606.16617 · PDF

  11. 11

    VeriGraph: Towards Verifiable Data-Analytic Agents

    Jiajie Jin, Zhao Yang, Wenle Liao, Yuyang Hu, Guanting Dong, Xiaoxi Li, Yutao Zhu, Zhicheng Dou

    cs.CL · cs.AI

    LLM-based agents have demonstrated strong capabilities in data-intensive analytical tasks, yet their outputs are rarely verifiable: a reliance on linear text trajectories makes their reasoning difficult to audit. In particular, deterministic computations over raw data and semantic deductions over natural-language claims are often entangled in an unstructured stream, leaving numerical conclusions hard to reproduce and qualitative judgments...

    arxiv.org/abs/2606.16603 · PDF

  12. 12

    Fast When, Careful Who: Dual-Process Multiparty Turn-Taking with Diffusion Augmentation

    Rutherford A. Patamia, Ming Liu, Wei Luo, Favour Ekong, Akan Cosgun

    cs.CL · cs.AI

    Reliable turn-taking is essential for spoken dialogue systems. However, most existing methods are designed for two-speaker interaction and struggle with realistic multiparty audio containing overlap and rapid speaker changes. We study multiparty turn-taking on the VoxConverse dataset and propose an audio-only two-stage pipeline that separates when to trigger a turn boundary from whether the floor is actually transferring. A fast trigger scans...

    arxiv.org/abs/2606.16568 · PDF

  13. 13

    REFLEX: Reflective Evolution from LLM Experience

    Pan Wang

    cs.CL · cs.LG

    Large multimodal language models (LLMs) have emerged as powerful tools for guiding evolutionary search toward interpretable programmatic policies. However, existing frameworks rely on a monolithic model call to simultaneously interpret visual behavioral evidence and synthesize corrective code. This diagnosis-repair entanglement creates an opaque feedback loop, obscuring the rationale behind mutations and preventing the retention of...

    arxiv.org/abs/2606.16496 · PDF

  14. 14

    Lost at the End: Primacy Bias in Multimodal Retrieval-Augmented Question Answering

    Jieyuan Liu, Jianyang Gu, Shijie Chen, Jefferson Chen, Zhen Wang

    cs.CL · cs.AI · cs.CV

    Knowledge-based visual question answering (KB-VQA) lets vision-language systems answer questions that exceed their parametric knowledge by conditioning a reader on passages retrieved from a Wikipedia-scale knowledge base. In pure-text long-context LLMs, retrieved-context use follows the U-shaped "lost-in-the-middle" effect of Liu et al. (2024): information at the start and end of context is used, the middle is lost. Whether this transfers to...

    arxiv.org/abs/2606.16494 · PDF

  15. 15

    ACCORD: Action-Conditioned Contextual Grounding for Language Agents

    Lai Jiang, Cheng Qian, Zhenhailong Wang, Pan Lu, Heng Ji, Hao Peng

    cs.CL · cs.AI

    User instructions are often underspecified because humans rely on implicit assumptions about the surrounding environment. For large language model (LLM) agents operating in information-rich digital and physical environments, these assumptions cannot be inferred from the instruction alone; they must be recovered from the current state of tools, data, interfaces, and observations. Effective execution therefore requires agents to identify...

    arxiv.org/abs/2606.16432 · PDF

  16. 16

    LectūraAgents: A Multi-Agent Framework for Adaptive Personalized AI-Assisted Learning and Embodied Teaching

    Jaward Sesay, Yue Yu, Siwei Dong, Yemin Shi, Guangyao Chen, Börje F. Karlsson

    cs.CL · cs.AI · cs.HC

    Effective personalized AI-assisted learning demands systems that can not only generate accurate learner-specific educational materials, but also dynamically adapt their instruction to diverse learners. However, existing educational agents have primarily focused on lecture content automation and simulations, which often fall short of modelling multimodal and embodied instructional methods tailored for the individual learner. To this end, we...

    arxiv.org/abs/2606.16428 · PDF

  17. 17

    A Mechanistic Understanding of Pronoun Fidelity in LLMs

    Katharina Trinley, Jesujoba O. Alabi, Dietrich Klakow, Vagrant Gautam

    cs.CL · cs.LG

    Faithful and robust pronoun use is important for fair and coherent generations, yet large language models largely fail when multiple referents use different pronouns. To study the interplay of reasoning, repetition, and bias in this task, prior work relies exclusively on behavioural approaches, which may not reflect a model's internal workings. Therefore, we provide a mechanistic, model-internal perspective on pronoun fidelity, testing...

    arxiv.org/abs/2606.16407 · PDF

  18. 18

    Evaluating LLM Personalization via Semantic Constraint Verification

    Xuran Li, Guanqin Zhang, Imran Razzak, Hakim Hacid, Eleanna Kafeza, Hao Xue, Flora D. Salim

    cs.CL · cs.LG

    Current evaluation paradigms for Large Language Model (LLM) personalization rely heavily on brittle surface-matching metrics or computationally expensive LLM-as-a-judge protocols, both of which lack interpretability. To address these limitations, we introduce Natural Language Inference Constraint Verification (NLICV), a scalable, semantically invariant framework that maps sentence meanings to truth-condition sets to verify personalization...

    arxiv.org/abs/2606.16368 · PDF

  19. 19

    Tyler: Typed Latent Reasoning for Language Models -- When to Think, What to Compute, and How Much to Allocate

    Hanyu Lin, Min Cai, Jiawei Wen, Haodi Zhang

    cs.CL · cs.AI

    Chain-of-thought (CoT) prompting improves reasoning in large language models (LLMs) by externalizing intermediate computation as discrete text tokens, but this textual interface also introduces redundancy and inference overhead. Latent reasoning offers a promising alternative by carrying part of the computation in continuous representations. However, existing methods typically predefine when latent computation is invoked and how it is...

    arxiv.org/abs/2606.16360 · 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.