Search: “arxiv-cs-cl”

When Updating Stops Being Learning: Rethinking LLM Self-Evolution via learnable information gain

arXiv:2609.36535v1 Announce Type: new Abstract: Self-evolution lets large language models (LLMs) improve iteratively using their own generated data, but often suffers from self-evolution degeneration:

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Model Releases

Which papyrus HTR is good enough? Character-error-rate tolerance of four papyrological tasks on Greek texts

DGX agent

arXiv:2609.37755v1 Announce Type: new Abstract: Purpose: Most Greek papyri remain unpublished and undigitised; a handwritten text recognition (HTR) pipeline that transcribes them automatically would l

model-releasesarxiv-cs-cl

30 Sept 2026

Who Warmed the Archives? LLMs Overestimate Historical Warmth

arXiv:2609.37499v1 Announce Type: new Abstract: Historical archives are an under-used source for extending the instrumental climate record backward in time, and LLMs offer a way to extract the indices

Your Benchmark Is Not Saturated: Reviving Multiple-Choice Evaluation with Answer Pooling

arXiv:2609.37494v1 Announce Type: new Abstract: Multiple-choice benchmarks are cheap to grade and are running out of room, and the standard remedy, writing harder items, is slow and repeated for every

Zero-shot Dependency Parsing with Unsupervised Cross-Lingual Bootstrapping

arXiv:2609.37883v1 Announce Type: new Abstract: Pre-trained language models (PLMs) with encoder-based architectures have shown impressive capabilities in zero-shot cross-lingual transfer for various l

A Benchmark Framework for Screening Automation in Systematic Reviews

arXiv:2609.30298v1 Announce Type: new Abstract: Systematic reviews (SR) are essential for evidence-based research, but their screening phase is highly time-consuming and labor-intensive. Large languag

AcoustiClaim: A Numeric Claim Benchmark with Instrument Ground Truth

arXiv:2609.30483v1 Announce Type: cross Abstract: Audio language models state numbers for acoustic quantities, and neither human opinion nor a judge model says whether such a number is true of the sig

All In Good Time: Causality-Aware Framework for LLM-Based Simultaneous Speech-to-Speech Translation

arXiv:2609.30416v1 Announce Type: new Abstract: Large Language Models (LLMs) have shown strong performance in low-resource offline translation; however, extending them to simultaneous speech-to-speech

Asymmetric Classifier-Free Guidance for Target-Speaker ASR

arXiv:2609.30476v1 Announce Type: cross Abstract: Target-speaker automatic speech recognition (TS-ASR) must identify and transcribe a desired speaker under varying overlap and noise conditions. These

Auditing and Repairing LLM-as-Judge Failures in a Production Text-to-SQL Pipeline

arXiv:2609.30290v1 Announce Type: new Abstract: Production text-to-SQL pipelines often end with an LLM-as-judge whose agreement with human annotators has never actually been measured. When we checked

Beyond Mean Attention: Diversity-Aware, Layer-Wise Scoring for KV Cache Eviction

arXiv:2609.30738v1 Announce Type: new Abstract: KV cache eviction methods such as SnapKV and PyramidKV rank tokens solely by mean attention over a small observation window. We study a unified score, m

CG-Probes: Recovering Guardrail Directions from Patient Query Embeddings

arXiv:2609.31062v1 Announce Type: new Abstract: Patient-facing AI assistants promise valuable support to patients, but incoming queries can pose medical risks. To create guardrails, we work with oncol

Coupled Usage-Sense Processes: Temporal and Attributable Lexical Semantic Change

arXiv:2609.30974v1 Announce Type: new Abstract: Lexical semantic change is usually summarized by a scalar distance between independently sampled period distributions. This measures how much a word cha

Cross-Backend QIEO: Universal Runtime Portability across OpenMP5, CUDA, HIP, and Multi-Language Interfaces

arXiv:2609.30914v1 Announce Type: cross Abstract: Quantum-inspired algorithms emulate quantum mechanical principles, such as, superposition, interference, and probabilistic amplitude evolution, on cla

Do we need to answer that question? Salience and Answerability of Potential Questions in Naturalistic Dialogue

arXiv:2609.31130v1 Announce Type: new Abstract: We empirically investigate Question Under Discussion based modelling in naturalistic dialogue by studying whether the salience of generated potential qu

Does Understanding Inform Generation in Unified Multimodal Models? From Analysis to Path Forward

arXiv:2511.20561v3 Announce Type: replace-cross Abstract: Recent years have witnessed significant progress in Unified Multimodal Models, yet a fundamental question remains: Does understanding truly in

Don't CLAP: Are Music-Text Models Bag-of-Words?

arXiv:2609.30540v1 Announce Type: cross Abstract: Text-to-music systems are assessed on audio quality and on how faithfully the music follows its prompt, and the CLAP score, the cosine similarity betw

Effects of Transcript Compression on LLM-based Medical Misinformation Detection in Japanese YouTube Videos

arXiv:2609.30882v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used to assess long-form medical videos, but their effectiveness may depend on whether transcripts are pro

Enhancing Assessment of Self-Consistency in LLM Explanations using Perturbation Strength

arXiv:2609.30849v1 Announce Type: new Abstract: Prior work has examined the self-consistency of LLM-generated explanations using surface-level perturbation methods. However, the strength of these pert

Epstein Files Engine: Agentic Search for Investigative Journalism

arXiv:2609.30611v1 Announce Type: cross Abstract: On Jan. 30, 2026, the U.S. Department of Justice released a mixed-media collection concerning Jeffrey Epstein, including about three million pages of

Evaluating Sycophancy in Chinese Large Language Models on Factual Questions Derived from Online Search Queries

arXiv:2609.30986v1 Announce Type: new Abstract: As large language models increasingly mediate information access, factually accurate and independent answers are critical. However, these models can exh

Evidence-Grounded Auditing of Identification Assumptions in Climate-Policy Causal Evaluations

arXiv:2609.30867v1 Announce Type: new Abstract: Difference-in-differences (DID) studies are widely used to evaluate climate policy, but assessing the evidence supporting their identification assumptio

FAVoR: Measuring and Mitigating Author-Style Homogenization in Federated Personalized Generation

arXiv:2609.30968v1 Announce Type: new Abstract: Large language models are increasingly used as personalized writing assistants, but adapting a model across many authors can compromise individual writi

Feeding BabyLMs Macaroni: Code-Switching Curricula Cause Cross-Lingual Convergence

arXiv:2609.30535v1 Announce Type: new Abstract: Children in multilingual communities often code-switch, using multiple languages in a single utterance. Can we induce cross-lingual alignment in languag

From annotation to reasoning: Culture in language models

arXiv:2609.30897v1 Announce Type: new Abstract: How should we evaluate language models when more than one interpretation can be right? Cultural benchmarks often test factual knowledge, agreement with

From ASR to ASP: Evaluating Prompt Attack Vulnerabilities Against Open-Source LLMs

arXiv:2505.14368v3 Announce Type: replace-cross Abstract: Recent studies demonstrate that Large Language Models (LLMs) are vulnerable to attacks that generate harmful or sensitive outputs. As open-sou

Hardware

I-Parakeet: Integer-Only Conformer ASR on Mobile NPU

arXiv:2609.30846v1 Announce Type: new Abstract: In this paper, we propose I-Parakeet, an integer-only implementation of NVIDIA's Parakeet-CTC (0.6B parameters) that runs on a smartphone NPU without an

hardwarearxiv-cs-cl

Identifying Scientists on X

arXiv:2609.31264v1 Announce Type: new Abstract: With the growing importance of science-related discourse on the Web and the erosion of the classical knowledge order, it is important to identify differ

Inference-Time Target Speaker Unlearning in LLM-Based Automatic Speech Recognition

arXiv:2609.30439v1 Announce Type: new Abstract: We introduce target-speaker unlearning ASR (TSU-ASR) task in a fully end-to-end framework for multi-speaker ASR and diarization. Given a multi-speaker u

Hardware

KuaFu: Compressing Long User Behavior into Understanding at Billion Scale

arXiv:2609.31045v1 Announce Type: cross Abstract: Conversational agents, generative recommenders, and personalized advertising all rest on one capability: understanding each user from raw behavior. Pr

hardwarearxiv-cs-cl

Learning Natural Conversational Behavior in Tandem Speech-to-Speech Models with Randomized Guidance

arXiv:2609.30773v1 Announce Type: new Abstract: Tandem speech-to-speech architectures couple a responsive speech frontend with an asynchronous text backend. In KAME, a large language model (LLM) serve

Local Ai

LocUS: Head Selection and Subspace Projection for Targeted Activation Steering

arXiv:2609.31122v1 Announce Type: new Abstract: Activation steering is a powerful training-free paradigm for controlling large language models at inference time. However, standard approaches estimate

local-aiarxiv-cs-cl

Manifold Projection and Iterative Autoencoder Refinement for Masked Language Modeling

arXiv:2609.30288v1 Announce Type: new Abstract: In Transformer-based masked language models, attention is the primary mechanism for context mixing, but there are other ways to mix data across tokens.

MexHat: A Dataset for Hate Speech Detection in Mexican Spanish Videos

arXiv:2609.31553v1 Announce Type: new Abstract: Ensuring online safety through content monitoring had raised Hate Speech Detection as a crucial task to be addressed. By essence the task demands the ca

Modeling Student Sensemaking with LLMs and Knowledge-Graph-Guided Inference

arXiv:2609.31046v1 Announce Type: new Abstract: Collaborative science learning requires nuanced interpretation of student dialogue to characterize how learners identify knowledge gaps, build explanati

Muslim: A Deployed Arabic Voice AI Platform for Grounded Islamic Knowledge

arXiv:2609.31511v1 Announce Type: new Abstract: We present Muslim, a production Arabic voice AI platform serving grounded, sourced Islamic knowledge to real users. Beyond a real-time voice pipeline (N

Not All Memories Are Equal: Hierarchical Collaborative Memory for Validity-Aware Retrieval in LLM Agents

arXiv:2609.30289v1 Announce Type: new Abstract: In team collaboration scenarios, memory is heterogeneous and continually evolving. Team memories capture collective decisions, protocols, and current co

PALM: Point-in-Time Adaptation for Financial Language Models

arXiv:2609.30316v1 Announce Type: cross Abstract: Language models used in financial backtests suffer from look-ahead bias, as a model trained on text published after the study period has already obser

Agents

PIA: A Personal Intelligence Agent Turning Health Conversations into Records and Records into Understanding

arXiv:2609.31255v1 Announce Type: new Abstract: General-purpose agent memory summarizes conversations: it extracts salient snippets, embeds them, and retrieves the top-k into the prompt. A health agen

agentsarxiv-cs-cl

Prompt Injection Detection for Email Agents Through Attack Chain Modeling

arXiv:2609.30657v1 Announce Type: cross Abstract: Large language model email assistants are particularly vulnerable to indirect prompt injection because untrusted email content can be retrieved into t

Quantizing Looped Transformers: Feedback Exposure and Calibration Blindness

arXiv:2609.30820v1 Announce Type: cross Abstract: Looped transformers reuse weights across recurrence steps, making low-bit quantization especially attractive. We identify two distinct failure modes o

RAZOR: Pruning Replaceable Experts in LLMs

arXiv:2609.30465v1 Announce Type: cross Abstract: Mixture-of-experts (MoE) models activate few experts per token but store the full expert pool. Expert pruning reduces this storage burden; at a fixed

Applications

REALMS: An AI-Assistant Conversational System for Real-Time Exact Audience Sizing over High-Dimensional Nested Profiles

arXiv:2609.30547v1 Announce Type: new Abstract: Audience sizing is a critical component of digital marketing. It enables precise resource allocation, campaign planning, and performance optimization. T

applicationsarxiv-cs-cl

Recursive Self-Improvement via On-Policy Distillation for Reasoning

arXiv:2609.30652v1 Announce Type: new Abstract: On-policy distillation (OPD) trains a student model by having it generate trajectories, then matching its next-token predictions with an external teache

RupeeBias: Auditing Demographic Bias in Indian Economic Guidance from Large Language Models

arXiv:2609.31245v1 Announce Type: new Abstract: Individuals turn to large language models (LLMs) for guidance across a wide range of economic tasks, from comparing loan options and planning savings to

SEA-CLIP-Tiny: Efficient Multilingual Text-Vision Embedding for Southeast Asian Languages

arXiv:2609.30739v1 Announce Type: new Abstract: Multilingual text-vision embedding models are essential for cross-lingual image-text retrieval, but Southeast Asian languages remain poorly supported du

Stale-Document Poisoning: When Outdated Retrieval Overrides Correct Model Answers

arXiv:2609.31342v1 Announce Type: new Abstract: Retrieval-augmented generation (RAG) is often used to address outdated knowledge by providing external evidence. But retrieval helps only when that evid

Local Ai

Statistical Foundations for a Google Play User-Review Sentiment Index: Signal Fusion, Shrinkage, Distributional Validation, and Dynamic Smoothing

arXiv:2609.31513v1 Announce Type: cross Abstract: We develop a statistically explicit sentiment index for Google Play user reviews and establish the mathematical results supporting its construction. N

local-aiarxiv-cs-cl