cs.CL · 2026-07-02 · No. 41

Computation and Language, 2026-07-02.

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

01 — The papers

18 entries
  1. 01

    Measuring the Gap Between Human and LLM Research Ideas

    Ziyu Chen, Yilun Zhao, Arman Cohan

    cs.CL · cs.AI

    LLMs are increasingly used to brainstorm research ideas, but existing evaluations mostly judge individual ideas by novelty, feasibility, or expert preference. We instead ask: how far are current LLM-generated ideas from human researchers? To characterize this gap, we build a large-scale evaluation framework for ideation from high-quality human research papers. For each paper, we reverse-engineer a small set of closely related prior works that...

    arxiv.org/abs/2607.01233 · PDF

  2. 02

    The State-Prediction Separation Hypothesis

    Giovanni Monea, Nathan Godey, Kianté Brantley, Yoav Artzi

    cs.CL · cs.AI · cs.LG

    Transformers use the same forward computation stream to both predict the next token and store useful state for future token predictions. We formulate the \emph{state-prediction separation hypothesis}: disentangling the two roles yields better language modeling performance. We design a Transformer variant that uses two computation streams to separate the two functions, and conduct pretraining experiments across various scales. Our experiments...

    arxiv.org/abs/2607.01218 · PDF

  3. 03

    Distill to Detect: Exposing Stealth Biases in LLMs through Cartridge Distillation

    Shayan Talaei, Abhinav Chinta, Devvrit Khatri, Amin Karbasi, Azalia Mirhoseini, Amin Saberi

    cs.CL · cs.AI · cs.LG

    Language models deployed in high-stakes roles can potentially favor certain entities, brands, or viewpoints, steering user decisions at scale. Such preferential biases can be introduced by any actor in the model's supply chain and are most dangerous when the model reveals its preference only on the relevant topic while behaving identically to its unmodified base on all other inputs. Recent work has shown that these biases can transfer through...

    arxiv.org/abs/2607.01208 · PDF

  4. 04

    Adversarial Pragmatics for AI Safety Evaluation: A Benchmark for Instruction Conflict, Embedded Commands, and Policy Ambiguity

    Brett Reynolds

    cs.CL · cs.AI · cs.SE

    Safety evaluations for language models increasingly depend on judgments about ambiguous natural-language behaviour: whether a model has followed an instruction, refused appropriately, complied with a policy, resisted an embedded command, or misreported progress in an agentic task. Existing benchmarks often compress these distinctions into pass/fail labels, obscuring whether failures arise from capability limits, policy ambiguity, instruction...

    arxiv.org/abs/2607.01153 · PDF

  5. 05

    Towards Developing a Multimodal Chat Assistant for University Stakeholders: RAG-based Approach

    Md Abu Hanif Shaikh, Abdullah Al Shafi

    cs.CL · cs.AI

    University stakeholders often face difficulties in accessing timely and reliable information, especially in developing countries, where there are very few intelligent support systems. Existing rule-based chatbots are unable to handle complex, domain-specific queries and are not well-equipped to adapt to evolving institutional policies. As a fill-in-the-gap solution, we present the multimodal university chatbot with retrieval-augmented...

    arxiv.org/abs/2607.01115 · PDF

  6. 06

    Message Passing Enables Efficient Reasoning

    Xuecheng Liu, Daman Arora, Gokul Swamy, Andrea Zanette

    cs.CL · cs.LG

    While inference-time scaling has improved the reasoning abilities of large language models (LLMs), the need to generate long chains-of-thought (CoTs) is a computational bottleneck. Thus, in contrast to sequential scaling methods like CoT, recent parallel scaling techniques instead use fork and join (FJ) primitives to divide work across multiple LLM threads. However, in the fork-join paradigm, threads are typically transient and do not...

    arxiv.org/abs/2607.01077 · PDF

  7. 07

    Behavior-Adaptive Conversational Agents: Toward a Fluid Personality Framework

    Hasibur Rahman, Smit Desai

    cs.CL · cs.AI · cs.HC

    Large language model (LLM)-based conversational agents (CAs) are now ubiquitous, creating new opportunities for AI-mediated behavior change. Their capacity to project nuanced personalities and adopt diverse metaphorical roles raises a design question: how should an agent's persona and personality be calibrated to the moment? Recent evidence suggests that (i) moderate personality expression outperforms low or high extremes on trust, enjoyment,...

    arxiv.org/abs/2607.01034 · PDF

  8. 08

    Reading Order Inference for Complex Document Layouts

    Iddo Hakim, Sharva Gogawale, Omer Ventura, Gal Grudka, Daria Vasyutinsky-Shapira, Berat Kurar-Barakat, Nachum Dershowitz

    cs.CL · cs.AI · cs.CV · cs.DL

    Reading order inference remains a critical bottleneck in the digitization of complex historical manuscripts, where pages contain multiple spatially interleaved reading streams, the canonical example being the Glossa Ordinaria layout, in which a central text is surrounded by commentaries that wrap around it in non-rectangular, non-convex regions. We present a training-free, graph-based framework: each OCR text line becomes a node in a directed...

    arxiv.org/abs/2607.01018 · PDF

  9. 09

    Logit-Contribution Scoring Identifies Non-Literal Retrieval Heads

    Aryo Pradipta Gema, Beatrice Alex, Pasquale Minervini

    cs.CL · cs.AI · cs.LG

    In long-context use, large language models frequently synthesize answers from the meaning of a relevant context span rather than literally copy-pasting them. Identifying which attention heads perform this synthesis matters for interpreting long-context model behavior. Yet existing detectors miss these heads by construction: they reward heads whose attended token matches the generated token, a literal-copy criterion that captures where a head...

    arxiv.org/abs/2607.01002 · PDF

  10. 10

    From Personas to Plot: Character-Grounded Multi-Agent Story Generation for Long-Form Narratives

    Aayush Aluru, Chloe Ho, Muhammad Hammouri, Kerry Luo, Myra Malik, Ryan Lagasse, Arjun Bahuguna, Vasu Sharma

    cs.CL · cs.AI · cs.MA

    Although large language models (LLMs) have demonstrated impressive creative fiction generation, they struggle to maintain narrative consistency and coherent plot lines in long-form stories. In this work, we introduce a unified framework for long-form narrative generation and verification. MAGNET, a multi-agent goal-driven narrative engine for storytelling, generates stories with persona-grounded character agents that propose actions based on...

    arxiv.org/abs/2607.00918 · PDF

  11. 11

    CAT: Confidence-Adaptive Thinking for Efficient Reasoning of Large Reasoning Models

    Qizhi Jiang, Shuo Wang, Pei Ke, Yuhang Song, Ke Qin

    cs.CL · cs.AI

    Large Reasoning Models (LRMs) have achieved remarkable success on complex tasks by leveraging long chain-of-thought (CoT) trajectories, yet they frequently exhibit overthinking on simple queries, resulting in significant token overhead and reduced inference efficiency. However, existing compression methods predominantly apply uniform length reduction or rely on coarse-grained difficulty estimation, often leading to performance degradation on...

    arxiv.org/abs/2607.00862 · PDF

  12. 12

    Recovering Input Text from Hidden States: Study of Gradient-Based Inversion of Decoder-Only Language Models

    Mikołaj Słowikowski, Maciej Witold Majewski

    cs.CL · cs.AI

    This work studies the hidden-state inversion problem: recovering the original input token sequence of a decoder-only language model from its last-layer hidden states. Rather than treating inversion as a one-shot reconstruction, we study it as a continuous embedding-space optimisation in which a soft proxy is driven towards the leaked target without any hard-token projection during the search, and a token is committed only once, at the end of...

    arxiv.org/abs/2607.00852 · PDF

  13. 13

    Self-conditioned Flow Map Language Models via Fixed-point Flows

    Jaehoon Yoo, Wonjung Kim, Floor Eijkelboom, Chanhyuk Lee, Nicholas M. Boffi, Seunghoon Hong, Jinwoo Kim

    cs.CL · cs.AI

    Self-conditioning is a core technique that enhances continuous flow-based language models, where the model learns to denoise generated text by conditioning on its own denoising estimate. While empirically successful, its performance improvements are poorly understood. Moreover, there is growing interest in the use of few-step generators based on flow maps, for which how to leverage self-conditioning is unclear. Here, we show that flow...

    arxiv.org/abs/2607.00714 · PDF

  14. 14

    Faithful by Definition: Emotion Analysis via Natural Semantic Metalanguage Explications

    Frank Xing, Erik Cambria

    cs.CL · cs.AI

    Explanations for emotion classifiers are usually produced post hoc, with no guarantee that they reflect the computation behind the label. We present an explication interface for event-based emotion analysis. A parser maps the input text to an explication, a short script in the closed vocabulary of Natural Semantic Metalanguage organized into twelve typed slots, and a fixed decision list of rules transcribed from published semantic definitions...

    arxiv.org/abs/2607.00661 · PDF

  15. 15

    Auditing Forgetting in Limited Memory Language Models

    Arya Raeesi, Hanna Roed

    cs.CL · cs.AI · cs.LG

    Limited Memory Language Models (LMLMs) externalize factual knowledge to a database to enable deletion-based unlearning without retraining. Existing evaluations measure post-deletion correctness in aggregate and cannot tell whether a deleted fact persists through residual parametric memory, alternative retrieval paths, or near-neighbor retrieval artifacts. We propose a causal auditing framework that holds the model fixed and varies the...

    arxiv.org/abs/2607.00605 · PDF

  16. 16

    BaseRT: Best-in-Class LLM Inference on Apple Silicon via Native Metal

    Prabod Rathnayaka, Fabian Waschkowski, Lukas Wesemann

    cs.CL · cs.AI · cs.PF

    We present BaseRT, a native Metal inference runtime for large language models (LLMs) on Apple Silicon, and report the highest inference throughput on this hardware to date. Existing runtimes, including llama.cpp and MLX-based frameworks, incur overhead from abstractions not designed for Metal's execution model or Apple Silicon's unified memory topology. By building natively on Metal with chip-specific kernel fusion, unified memory-aware...

    arxiv.org/abs/2607.00501 · PDF

  17. 17

    A Mechanistic View of Authority Hierarchy in LLM Sycophancy

    Emil Joswin, Srujananjali Medicherla, Priyanka Mary Mammen

    cs.CL · cs.LG

    Authority bias poses a critical safety concern in language models: models systematically prioritize social cues from authority figures over factual consistency, swaying their answers based on source credibility rather than evidence. We mechanistically investigate this phenomenon using a controlled medical QA setting, where hints suggesting incorrect answers are attributed to personas of varying expertise. Across Llama-3.1-8B, Qwen3-8B, and...

    arxiv.org/abs/2607.00415 · PDF

  18. 18

    DiscoLoop: Looping Discrete Embeddings and Continuous Hidden States for Multi-hop Reasoning

    Hengyu Fu, Tianyu Guo, Zixuan Wang, Hanlin Zhu, Jason D. Lee, Jiantao Jiao, Stuart Russell, Song Mei

    cs.CL · cs.AI · cs.LG

    Large language models achieve strong performance on many reasoning tasks when allowed to externalize intermediate steps as Chain-of-Thought (CoT). However, many questions require the model to internalize the multi-step reasoning within a single forward pass before generating the answer. We study this challenge through two-hop reasoning, a representative task where the model must compose multiple pieces of parametric knowledge within a single...

    arxiv.org/abs/2607.00341 · 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.