Search: “arxiv-cs-cl”

Predicting and Repairing Merge Collapse in Large Language Models

arXiv:2610.03199v1 Announce Type: cross Abstract: Large language models fine-tuned from a shared base can be merged by averaging their task vectors, but some merges collapse far below the base model,

Hardware

Human

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Probe the Harness: Setup Checks for Stale-Data RL Comparisons in Language Models

DGX agent

arXiv:2610.02911v1 Announce Type: cross Abstract: Methods for training language models on stale samples are judged by comparisons against importance-corrected baselines. We show that details of the ex

hardwarearxiv-cs-cl

5 Oct 2026

Recursive Self-Improvement in Unified Multimodal Models

arXiv:2610.03002v1 Announce Type: new Abstract: Unified multimodal models (UMMs) understand and generate both text and images, which lets a model produce its own training data. Existing self-improveme

Applications

ReSCUE: Re-translation with Sentence Commitment for Unsegmented Long-Form Simultaneous Sign Language Translation

arXiv:2610.03022v1 Announce Type: cross Abstract: Simultaneous Sign Language Translation (SLT) is critical for real-time communication, yet existing methods remain largely confined to sentence-level,

applicationsarxiv-cs-cl

RMCW: A Deletion-Robust Watermark Based on Reed--Muller Codes for Language Models

arXiv:2610.02817v1 Announce Type: cross Abstract: Large Language Model (LLM) watermarking provides a lightweight mechanism for identifying text generated by a specific model, but its robustness remain

RT-SFT: Text Style Transfer from Non-Parallel Corpora by Roundtrip Translation

arXiv:2602.15013v2 Announce Type: replace Abstract: Text style transfer (TST) is naturally a supervised task - rewrite a sentence in a target style while preserving its meaning - yet the parallel corp

Safety

SecJev: Bringing Security Expertise to System One Decision Models

arXiv:2610.03073v1 Announce Type: cross Abstract: Security workflows need models that turn complex observations and explicit policies into decisions. System One models introduced by Jev return typed p

safetyarxiv-cs-cl

SEDIMA: Cross-Run Hierarchical Insight Memory for Evolutionary Search Agents

arXiv:2610.02361v1 Announce Type: cross Abstract: Large language model (LLM)-driven evolutionary search is a powerful paradigm for automated program and algorithm discovery, yet existing systems are l

Shrome at Touche: Soft-Vote Ensembling and Counter-Causal Augmentation for Causality Extraction

arXiv:2610.03268v1 Announce Type: new Abstract: Touche 2026 extends causality extraction to counter-causal claims: news sentences whose surface form appears causal but whose meaning denies the causati

Silent Dissent: LLM Agents That Yield to the Majority Still Represent Their Original Premise

arXiv:2610.02702v1 Announce Type: new Abstract: Multi-agent debate is increasingly used to reach consensus among LLM agents, yet agents often yield to a unanimous majority. When an agent changes its a

Single-Pass Uncertainty Heads for Claim-Level Hallucination Detection in Persian Medical Language Models

arXiv:2610.03482v1 Announce Type: new Abstract: Hallucination detection is particularly important for medical language models, but repeated-sampling approaches are expensive and existing uncertainty-h

Applications

Social bot detection in the age of ChatGPT: Challenges and opportunities

arXiv:2610.02386v1 Announce Type: cross Abstract: We present a comprehensive overview of the challenges and opportunities in social bot detection in the context of the rise of sophisticated AI-based c

applicationsarxiv-cs-cl

Agents

Source Preference in the Wild: How LLM Agents Favor Items by Source, and How to Reduce It

arXiv:2610.03195v1 Announce Type: new Abstract: As LLM agents decide on users' behalf which product to buy, which hotel to book, or which paper to cite, a preference for items from certain sources (th

agentsarxiv-cs-cl

StanceEval 2026: The Second Stance Detection Shared Task

arXiv:2610.03215v1 Announce Type: new Abstract: StanceEval 2026 is the second edition of the StanceEval shared task series on stance detection in Arabic social media text. Stance detection aims to ide

Safety

Structured Composition of Verifiable Atomic Insights for Table-to-Report Generation

arXiv:2610.03525v1 Announce Type: new Abstract: Table-to-report generation refers to the task of automatically generating article-level analyt- ical reports from relational tables and is an essential

safetyarxiv-cs-cl

Local Ai

Text-Centric Post-Training for Omni-Modal Reasoning

arXiv:2610.02819v1 Announce Type: new Abstract: Improving joint audio-visual reasoning in Omni Large Language Models typically incurs substantial data construction and training costs. Our diagnostics

local-aiarxiv-cs-cl

The Geometry of Knowledge Accessibility in Large Language Models

arXiv:2610.03052v1 Announce Type: new Abstract: Large language models (LLMs) contain broad knowledge, but they cannot access all of it reliably. We study this problem through knowledge accessibility,

The Percept-V Challenge: Can Multimodal LLMs Crack Simple Perception Problems?

arXiv:2508.21143v4 Announce Type: replace Abstract: Cognitive science research treats visual perception, the ability to understand and make sense of a visual input, as one of the early developmental s

To Explore The Strange New World Beyond Data Distribution: System Behavior, Causality Tax, and Non-causal Base Model

arXiv:2610.02839v1 Announce Type: cross Abstract: We show that the causality of language models (LMs) may not be necessary nor optimal. This is the case when system behavior (denoted as S) is incorpor

To Jev or Not? Evaluating the Accuracy and Efficiency of Structured Decision Models for Hate-Speech Moderation

arXiv:2610.03324v1 Announce Type: new Abstract: The scale of online content makes hate-speech moderation challenging, while Large Language Models (LLMs) enable harmful material to be produced and adap

Trained Agentic Context Management

arXiv:2610.02404v1 Announce Type: cross Abstract: We study long context language models. Instead of training long context natively, or designing a long context harness, we train a model over the simpl

Understanding Clinical Cognitive Dialogues Using Large Language Models

arXiv:2609.34125v2 Announce Type: replace Abstract: In-person cognitive assessment is both a test and an interaction. Clinicians explain tasks, repair misunderstandings, and adapt to patient responses

Understanding Trajectory Heterogeneity in Federated World Model Learning

arXiv:2610.02957v1 Announce Type: cross Abstract: World models learn state evolution from trajectories, making access to temporal context a central training requirement. Federated learning can use dis

Unmasking Propaganda: A Comparative Analysis of Masked and Causal Language Models

arXiv:2610.03077v1 Announce Type: new Abstract: Propaganda detection is an essential task in natural language processing (NLP), particularly in the context of manipulative political communications. Ho

WakeKV: Reactive, Reversible KV Residency for Heads That Change Their Minds

arXiv:2610.02713v1 Announce Type: new Abstract: Most KV-cache compression methods classify attention heads once, either offline or during prefill, and keep this classification fixed throughout generat

World Embedding Benchmark

arXiv:2610.03632v1 Announce Type: cross Abstract: Physical fidelity has received increasing attention in world models and video generation, yet how video representations encode physical information re

Local Ai

Writerslogic at PAN 2026: Process over Content for Robust Detection under Domain Shift

arXiv:2610.03565v1 Announce Type: new Abstract: We describe the Writerslogic systems for three PAN at CLEF 2026 shared tasks (Reasoning Trajectory Detection, Voight-Kampff Generative AI Detection, and

local-aiarxiv-cs-cl

Writerslogic at the CLEF 2026 SimpleText Track: Multi-Candidate LLM Simplification and Stacked Complexity Spotting

arXiv:2610.03567v1 Announce Type: new Abstract: We describe the Writerslogic team's participation in the CLEF 2026 SimpleText shared task, addressing Task 1 (text simplification) and Task 2 (complexit

4MT-VLM: How Coarse Is a VLMs Cognitive Map?

arXiv:2609.39238v1 Announce Type: new Abstract: An agent that moves must recognise a place from a viewpoint it has never seen. We introduce 4MT-VLM, a dataset of procedurally generated landscapes, eac

A Missing Piece for Trustworthy AI Reviewers: From Benchmarking Rhetorical Robustness to SciCore Review

arXiv:2609.39027v1 Announce Type: new Abstract: AI reviewers can assign different judgments to manuscripts that report the same science in different wording, potentially rewarding rhetorical optimizat

A Tilted Bowl Is Not a Slippery Slope: Compressing Looped Models

arXiv:2609.39277v1 Announce Type: cross Abstract: Looped models reason by applying the same block of weights many times, so compressing that block saves memory traffic on every loop. Compressed looped

Tutorials

AdaGEPA: Adaptive Feedback Allocation for Reflective Prompt Optimization

arXiv:2609.39927v1 Announce Type: new Abstract: Prompt optimization improves the performance of language-model systems on downstream tasks by refining their prompts. Classical methods evaluate prompts

tutorialsarxiv-cs-cl

Anchor-ECC: Local Integrity Checking for Watermarked LLM Outputs via Error-Correcting Codes

arXiv:2609.38722v1 Announce Type: cross Abstract: LLM watermarking has become an effective approach to distinguishing AI-generated text from human-written text by embedding detectable patterns during

Applications

Anthropomorphism in the age of Large Language Models: An overview of potential risks and mitigations

arXiv:2609.38486v1 Announce Type: cross Abstract: Large Language Models (LLMs) and more broadly Artificial Intelligence (AI) systems are often described and understood in human-like terms, a phenomeno

applicationsarxiv-cs-cl

Argument Structure Prediction in Online Conversations: A Comparative Study of Modeling Paradigms and Task Architectures

arXiv:2609.39225v1 Announce Type: new Abstract: Argument structure prediction (ASP) constructs complete argument structures from discourse by identifying argumentative units and their relations. While

Atomic and Holistic LLM Judges for Reference-Grounded Support Labels: A Prompt-Controlled Comparison

arXiv:2603.28005v2 Announce Type: replace Abstract: When an LLM judge only has to assign a three-way support label to a candidate answer given a reference, does asking it to decompose the answer into

Agents

AutoDataBench: A Data-centric Testbed for Accelerating Auto Research

arXiv:2609.40097v1 Announce Type: new Abstract: Existing auto-research benchmarks often entangle multiple sources of improvement, including training frameworks, hyperparameters, compute budgets, and d

agentsarxiv-cs-cl

Better Supervision Is Nearby: Neighborhood On-Policy Self-Distillation

arXiv:2609.39687v1 Announce Type: new Abstract: On-policy self-distillation (OPSD) trains mathematical reasoning models using a privileged teacher that sees a reference solution and supervises student

Tutorials

Blackboard Intelligence Can Surpass Autoregressive on Globally Constrained Problems

arXiv:2609.38806v1 Announce Type: cross Abstract: Next-token prediction has driven remarkable progress in large language models, yet a growing body of evidence suggests that they can struggle on probl

tutorialsarxiv-cs-cl

Hardware

Bongard: Training Machine Intuition

arXiv:2609.39111v1 Announce Type: new Abstract: Human intelligence relies heavily on learned intuition: recognising patterns and judging situations without explicitly unfolding every intermediate step

hardwarearxiv-cs-cl

Agents

Can Terminal Agents Trust Their Own Verification? Diagnosing and Improving Self-Verification

arXiv:2609.38812v1 Announce Type: new Abstract: Terminal agents rely on self-verification to assess and correct their solutions as they solve tasks through interaction with command-line environments.

agentsarxiv-cs-cl

CATCH: A Controllable Analysis Testbed for Reward Hacking in Coding RL

arXiv:2609.39533v1 Announce Type: new Abstract: During reinforcement learning with verifiable rewards (RLVR), large language models (LLMs) can exploit loopholes in their environments to obtain high re

Cheap to Draw, Expensive to Trust: Certifying Test-Time Scaling Curves

arXiv:2609.40190v1 Announce Type: cross Abstract: Sampling several answers and keeping the one a verifier scores highest is one of the simplest ways to buy accuracy at test time. Its effect is reporte

Safety

Cognitive Enhancement: Rethinking the Necessity of Role-Playing for Large Language Models

arXiv:2609.39853v1 Announce Type: new Abstract: Role-playing prompting has become a popular yet simple technique for improving LLM reasoning and output quality. However, whether it consistently boosts

safetyarxiv-cs-cl

Compact Language, Complex Model Shifts: How and Where Ambiguity and Underspecification Affect LLMs

arXiv:2609.39572v1 Announce Type: new Abstract: We analyze how lexical ambiguity and underspecification affect language model training. We create artificial homonyms and artificial hypernyms as pseudo

Comparison of techniques for fine-tuning open-weight models for entity extraction from radiology reports

arXiv:2609.40236v1 Announce Type: new Abstract: Converting free-text radiology reports into structured labels supports cohort building, quality assurance, and monitoring of clinical imaging models, bu

Concept Subspaces Compute Beyond the Logit Lens: A Weights-Only Test for Locating Representations Upstream of Readout

arXiv:2609.39263v1 Announce Type: new Abstract: A concept subspace's effect on model behavior does not establish how it relates to the output readout. We introduce a two-sided geometric diagnostic tha

Conformal Factuality Control for Multi-Hop Retrieval-Augmented Generation

arXiv:2609.38222v1 Announce Type: new Abstract: Retrieval-augmented generation (RAG) can ground large language models in external evidence, but retrieved context does not guarantee that generated clai