Computation and Language

Papers filed under cs.CL on arXiv, each one already summarized by Paperlayer. Open any of them to read the summary beside the original PDF, with every point linked to the line, figure, or table it came from.

Search paper metadata (including unsummarized papers)

11,041 to 11,100 of 11,224

When Lower Privileges Suffice: Investigating Over-Privileged Tool Selection in LLM Agents
Kaiyue Yang, Yuyan Bu, Jingwei Yi +5
cs.SEcs.AIcs.CLarXiv:2606.20023v22026
Demystifying Training-Time Augmentation for Data-Constrained Language Model Pretraining
Michael K. Chen, Xikun Zhang, Fan Bai +2
cs.LGcs.AIcs.CLarXiv:2606.16246v22026
RoPE-Aware Bit Allocation for KV-Cache Quantization
Fengfeng Liang, Yuechen Zhang, Jiaya Jia
cs.LGcs.CLarXiv:2606.24033v12026
Remember When It Matters: Proactive Memory Agent for Long-Horizon Agents
Yifan Wu, Lizhu Zhang, Yuhang Zhou +5
cs.AIcs.CLarXiv:2607.08716v12026
Towards Mechanistically Understanding Why Memorized Knowledge Fails to Generalize in Large Language Model Finetuning
Lu Dai, Ziyang Rao, Yili Wang +3
cs.AIcs.CLarXiv:2607.08393v12026
A Sovereign, Open-Source Foundation Model for German and English
Soofi-Team, :, Benedikt Droste +30
cs.CLcs.AIcs.LGarXiv:2607.09424v32026
PAST-TIDE: Prototype-Anchored Statement Tuning with Topic-Invariant Normalization for Stance Detection
Md. Shakhoyat Rahman Shujon, MD Jahid Hasan Jim, Md. Milon Islam +2
cs.CLcs.LGarXiv:2607.04690v12026
Look Light, Think Heavy: What Multimodal Chain-of-Thought Reasoning Can and Cannot Do
Zhuoran Jin, Kejian Zhu, Hongbang Yuan +5
cs.CLcs.AIcs.CVarXiv:2606.22565v12026
Managing Procedural Memory in LLM Agents: Control, Adaptation, and Evaluation
Julia Belikova, Rauf Parchiev, Evgeny Egorov +4
cs.AIcs.CLcs.SEarXiv:2606.23127v12026
Constraint Tax in Open-Weight LLMs: An Empirical Study of Tool Calling Suppression Under Structured Output Constraints
Fangzheng Li, Aimin Zhang, Chen Lv
cs.CLarXiv:2606.25605v12026
Cognitive Episodes in LLM Reasoning Traces Enable Interpretable Human Item Difficulty Prediction
Chenguang Wang, Ming Li, Xinyue Zeng +4
cs.CLcs.AIcs.CYarXiv:2606.28186v32026
Measuring the Gap Between Human and LLM Research Ideas
Ziyu Chen, Yilun Zhao, Arman Cohan
cs.CLcs.AIarXiv:2607.01233v12026
VIBE: Voice-Induced open-ended Bias Evaluation for Large Audio-Language Models via Real-World Speech
Yi-Cheng Lin, Yusuke Hirota, Sung-Feng Huang +1
eess.AScs.CLcs.SDarXiv:2604.17248v22026
When More Sampling Hurts: The Modal Ceiling and Correlation Ceiling of Test-Time Scaling
Yong Yi Bay, Kathleen A. Yearick
cs.LGcs.AIcs.CLarXiv:2606.28661v12026
ENTRAP-VL: A Taxonomic Probe for Dual Contextual Entrainment in Vision-Language Models
Karan Goyal, Afreen Hossain, Debojyoti Das +1
cs.CVcs.AIcs.CLarXiv:2607.20092v12026
No Universal Signal Predicts Sample-Level LLM Regression under Version Updates
Jia Sheng, Yiwei Lu
cs.AIcs.CLcs.LGarXiv:2608.13607v12026
Asymmetric Discourse Homogenization and Shared Language Technology: Evidence from Reddit
Fengming Liu
cs.CYcs.CLcs.SIarXiv:2608.13674v12026
A Calibrated Test of Internal Action Maps: State Signals Without Global Affine Closure
Dekun Yang
cs.AIcs.CLcs.LGarXiv:2608.13626v12026
Computational Humor with Multimodal LLMs: Methods, Datasets, Evaluation, and Challenges
Tuo Liang, Zhe Hu, Disheng Liu +2
cs.CLcs.AIcs.MMarXiv:2607.19011v12026
Self-State Attacks on Self-Hosted AI Agents: How Far Can OS Defenses Go?
Yimeng Chen, Nathanaël Denis, Roberto Di Pietro +1
cs.CRcs.AIcs.CLarXiv:2607.17986v12026
Evidence Attribution in Visual Document Understanding without Coordinates or Region Labels
Zhuchenyang Liu, Yao Zhang, Yu Xiao
cs.CVcs.CLcs.IRarXiv:2607.24651v12026
Token-Level Off-Policy Learning for Faithful Generation Under Distribution Shift
Zitong Huang, Gustavo Lucas Carvalho, Deqing Fu +1
cs.CLcs.LGarXiv:2607.17524v12026
AI Tour Meeting: Group Travel Planning by LLM Agents
Daisuke Kikuta
cs.AIcs.CLcs.MAarXiv:2607.18806v12026
Train the Model, Not the Reader: Decodability Supervision for Verifiable Activation Explanations
Hiskias Dingeto
cs.AIcs.CLarXiv:2607.20379v12026
Temporal-Distance JEPA: Plan-Aware Representation Learning for Latent World Model Predictive Control
Jiaxin Bai, Jiaxuan Xiong
cs.CLcs.ROarXiv:2607.25337v22026
O-VAD: Industrial Video Anomaly Detection through Object-Centric Tracking and Reasoning
Mei Yuan, Qi Long, Qifeng Wu +5
cs.CVcs.AIcs.CLarXiv:2607.18142v12026
Where Quality Breaks in Compressed Short-Text Generation: Staged Bottleneck Localization
Alexey Gavrilov, Alan-Barsag Gazzaev, Sergey Muravyov
cs.CLcs.LGarXiv:2607.24176v12026
OpenForgeRL: Train Harness-native Agents in Any Environment
Xiao Yu, Baolin Peng, Ruize Xu +7
cs.AIcs.CLarXiv:2607.21557v32026
A Frozen 12B Beats Frontier Models on Verified Work: 100% Accuracy, 0 Tokens, Bit-Exact, Forever
Sietse Schelpe
cs.CLcs.AIcs.IRarXiv:2607.23806v12026
Fairness Pruning: Locating Demographic Bias in GLU-MLP Layers via Differential Activations
Pere Martra, Eugenio Martínez Cámara, Alfonso Ureña López
cs.CLcs.CYcs.LGarXiv:2607.28319v12026
H$^2$SD: Hybrid Hindsight Self-Distillation
Qiye Cai, Yichuan Ma, Peiji Li +6
cs.LGcs.CLarXiv:2607.18955v42026
LLM-as-a-Coach: Experiential Learning for Non-Verifiable Tasks
Tianzhu Ye, Li Dong, Guanheng Chen +4
cs.LGcs.CLarXiv:2607.18110v12026
Reading and Steering Representations of Materials-Science Mechanisms in an Open-Weight Language Model
Markus J. Buehler
cs.AIcond-mat.mes-hallcond-mat.mtrl-sciarXiv:2607.20058v12026
SecRespond: Benchmarking AI Agents for Real-World Post-Compromise Incident Response
Lehan Wang, Boli Chen, Ruixue Ding +7
cs.CRcs.AIcs.CLarXiv:2607.26791v12026
Tencent WorkBuddy Bench: A Multi-Domain Coding-Agent Benchmark with Contamination-Resistant Task Construction
Tencent WorkBuddy Bench Team, Siqi Cai, Shaopeng Chen +35
cs.CLcs.SEarXiv:2607.20911v12026
Leveraging External Knowledge for Historical Document Restoration via Retrieval-Augmented Large Language Models
Gabeen Kim, Kyeongpil Kang
cs.CLarXiv:2607.21936v12026
WorldCupArena: Fine-Grained Evaluation of Language Models and Deep-Research Agents on Football Forecasting
Zhaokai Wang, Tianlin Gui, Jiayuan Rao +3
cs.AIcs.CLcs.LGarXiv:2607.18084v12026
DocOps: A Verifiable Benchmark for Autonomous Agents in Complex Document Operations
Jiazhen Jiang, Boxi Cao, Lingyong Yan +6
cs.AIcs.CLcs.LGarXiv:2607.19865v12026
AutoIndex: Learning Representation Programs for Retrieval
Sam O'Nuallain, Nithya Rajkumar, Ramya Narayanasamy +3
cs.IRcs.AIcs.CLarXiv:2607.18603v12026
Would You Walk to the Car Wash? Revealing the Salience Bias of Large Language Models in Commonsense Reasoning
Zheng Wu, Chenhao Xue, Shijie Zheng +3
cs.CLarXiv:2607.28478v12026
Beyond Feeling Better: Capability-Sustaining Emotional Dialogue as a Longitudinal Research Paradigm
Ming Wang, Jiaqi Wu Young, Wenfang Wu +2
cs.CLcs.HCcs.SIarXiv:2607.27851v12026
Towards Robust Reinforcement Learning for Small-Scale Language Model Agents
Md Rezwanul Haque, Md. Milon Islam, Fakhri Karray
cs.AIcs.CLcs.LGarXiv:2607.25091v12026
DataPrep-Bench: Benchmarking LLMs as Training Data Preparators
Hao Liang, Qifeng Cai, Yibo Lin +11
cs.LGcs.CLarXiv:2607.20465v12026
Scaling Laws for Hypernetwork-Based Knowledge Injection in Large Language Models
Nischay Dhankhar, Dos Baha, Abulhair Saparov
cs.CLcs.LGarXiv:2607.19604v12026
Scaling Native Multimodal Pre-Training From Scratch
Haoyuan Wu, Aoqi Wu, Hai Wang +3
cs.CLcs.CVarXiv:2607.22043v12026
GPT-Red: Automated Red Teaming via Self-Play at Scale
Eric Wallace, Christopher A. Choquette-Choo, Nikhil Kandpal +15
cs.CRcs.AIcs.CLarXiv:2607.26115v12026
Filesystem-Based Memory for LLM Agents: Organization, Evolution, and Sustainability
Sizhe Zhou, Sheldon Yu, Hui Wei +8
cs.CLcs.AIarXiv:2607.26637v12026
SpecFirst: Behavioral Specification Elicitation as a First-Class Step in Agent-Based Program Synthesis from Scratch
Yihao Chen, Shi Chang, Feng Lin +4
cs.SEcs.CLarXiv:2607.27167v12026
Transcription Policy as a Latent Variable: Activating Controllable Verbatim ASR with Word-Level Timing
Laurin Wagner, Mario Zusag, Bernhard Thallinger
cs.CLarXiv:2607.18934v12026
Memory for Large Language Models
Sining Zhoubian, Dan Zhang, Evgeny Kharlamov +1
cs.CLarXiv:2607.25380v12026
Novel Claim or Déjà Vu? Rethinking "Contamination-Free'' Dynamic Evaluation for Multimodal Automated Fact-Checking
Haorui He, Xinwen Chen, Dacheng Wen +3
cs.CLcs.AIcs.MMarXiv:2607.23514v12026
Stale but Stable: Staleness-Adaptive Trust Regions for Stabilizing Asynchronous Reinforcement Learning
Junyao Yang, Yucheng Shi, Zongxia Li +6
cs.LGcs.CLarXiv:2607.18722v32026
Shieldstral
Antonia Calvi, Avinash Sooriyarachchi, Giada Pistilli +273
cs.CLcs.CVarXiv:2607.25857v22026
MindForge: Teaching Small Language Models Whole-Life-Cycle Software Engineering via Source-Free Program Synthesis
Yihao Chen, Shi Chang, Khaled Chawa +4
cs.SEcs.CLcs.LGarXiv:2607.27146v12026
AgentDebugX: An Open-Source Toolkit for Failure Observability, Attribution, and Recovery in LLM Agents
Kunlun Zhu, Xuyan Ye, Zhiguang Han +9
cs.AIcs.CLarXiv:2607.18754v12026
Envs-FORGE: Frontier-Optimized Reward-Grounded Environment Synthesis for Agent RL
Xiaojun Wu, Cehao Yang, Honghao Liu +7
cs.CLarXiv:2608.14312v12026
You Only Pass Once: Answering and Abstaining Together in a Single Forward Pass of a Frozen Language Model
Ziyang Luo, Zhongyao Chu, Xinjie He +4
cs.CLcs.LGarXiv:2608.14465v12026
Measuring Fairness in Large Audio Language Models via Semantic-Aware Bias Estimation
Zhe Liu
cs.CLcs.AIcs.SDarXiv:2608.13624v12026
Beyond Relevance-Centric Retrieval: Rubric-Oriented Document Set Selection and Ranking
Kailin Jiang, Lei Liu, Jian Xi +7
cs.CLarXiv:2607.19747v22026
OmegaUse-OfficeVal: Benchmarking LLM Agents on Long-Horizon Office-Suite Tasks with Economic Grounding
Jingbo Zhou, Yusai Zhao, Qi Bao +12
cs.AIcs.CLcs.HCarXiv:2607.27155v12026

Computation and Language

When Lower Privileges Suffice: Investigating Over-Privileged Tool Selection in LLM Agents

Demystifying Training-Time Augmentation for Data-Constrained Language Model Pretraining

RoPE-Aware Bit Allocation for KV-Cache Quantization

Remember When It Matters: Proactive Memory Agent for Long-Horizon Agents

Towards Mechanistically Understanding Why Memorized Knowledge Fails to Generalize in Large Language Model Finetuning

A Sovereign, Open-Source Foundation Model for German and English

PAST-TIDE: Prototype-Anchored Statement Tuning with Topic-Invariant Normalization for Stance Detection

Look Light, Think Heavy: What Multimodal Chain-of-Thought Reasoning Can and Cannot Do

Managing Procedural Memory in LLM Agents: Control, Adaptation, and Evaluation

Constraint Tax in Open-Weight LLMs: An Empirical Study of Tool Calling Suppression Under Structured Output Constraints

Cognitive Episodes in LLM Reasoning Traces Enable Interpretable Human Item Difficulty Prediction

Measuring the Gap Between Human and LLM Research Ideas

VIBE: Voice-Induced open-ended Bias Evaluation for Large Audio-Language Models via Real-World Speech

When More Sampling Hurts: The Modal Ceiling and Correlation Ceiling of Test-Time Scaling

ENTRAP-VL: A Taxonomic Probe for Dual Contextual Entrainment in Vision-Language Models

No Universal Signal Predicts Sample-Level LLM Regression under Version Updates

Asymmetric Discourse Homogenization and Shared Language Technology: Evidence from Reddit

A Calibrated Test of Internal Action Maps: State Signals Without Global Affine Closure

Computational Humor with Multimodal LLMs: Methods, Datasets, Evaluation, and Challenges

Self-State Attacks on Self-Hosted AI Agents: How Far Can OS Defenses Go?

Evidence Attribution in Visual Document Understanding without Coordinates or Region Labels

Token-Level Off-Policy Learning for Faithful Generation Under Distribution Shift

AI Tour Meeting: Group Travel Planning by LLM Agents

Train the Model, Not the Reader: Decodability Supervision for Verifiable Activation Explanations

Temporal-Distance JEPA: Plan-Aware Representation Learning for Latent World Model Predictive Control

O-VAD: Industrial Video Anomaly Detection through Object-Centric Tracking and Reasoning

Where Quality Breaks in Compressed Short-Text Generation: Staged Bottleneck Localization

OpenForgeRL: Train Harness-native Agents in Any Environment

A Frozen 12B Beats Frontier Models on Verified Work: 100% Accuracy, 0 Tokens, Bit-Exact, Forever

Fairness Pruning: Locating Demographic Bias in GLU-MLP Layers via Differential Activations

H$^2$SD: Hybrid Hindsight Self-Distillation

LLM-as-a-Coach: Experiential Learning for Non-Verifiable Tasks

Reading and Steering Representations of Materials-Science Mechanisms in an Open-Weight Language Model

SecRespond: Benchmarking AI Agents for Real-World Post-Compromise Incident Response

Tencent WorkBuddy Bench: A Multi-Domain Coding-Agent Benchmark with Contamination-Resistant Task Construction

Leveraging External Knowledge for Historical Document Restoration via Retrieval-Augmented Large Language Models

WorldCupArena: Fine-Grained Evaluation of Language Models and Deep-Research Agents on Football Forecasting

DocOps: A Verifiable Benchmark for Autonomous Agents in Complex Document Operations

AutoIndex: Learning Representation Programs for Retrieval

Would You Walk to the Car Wash? Revealing the Salience Bias of Large Language Models in Commonsense Reasoning

Beyond Feeling Better: Capability-Sustaining Emotional Dialogue as a Longitudinal Research Paradigm

Towards Robust Reinforcement Learning for Small-Scale Language Model Agents

DataPrep-Bench: Benchmarking LLMs as Training Data Preparators

Scaling Laws for Hypernetwork-Based Knowledge Injection in Large Language Models

Scaling Native Multimodal Pre-Training From Scratch

GPT-Red: Automated Red Teaming via Self-Play at Scale

Filesystem-Based Memory for LLM Agents: Organization, Evolution, and Sustainability

SpecFirst: Behavioral Specification Elicitation as a First-Class Step in Agent-Based Program Synthesis from Scratch

Transcription Policy as a Latent Variable: Activating Controllable Verbatim ASR with Word-Level Timing

Memory for Large Language Models

Novel Claim or Déjà Vu? Rethinking "Contamination-Free'' Dynamic Evaluation for Multimodal Automated Fact-Checking

Stale but Stable: Staleness-Adaptive Trust Regions for Stabilizing Asynchronous Reinforcement Learning

Shieldstral

MindForge: Teaching Small Language Models Whole-Life-Cycle Software Engineering via Source-Free Program Synthesis

AgentDebugX: An Open-Source Toolkit for Failure Observability, Attribution, and Recovery in LLM Agents

Envs-FORGE: Frontier-Optimized Reward-Grounded Environment Synthesis for Agent RL

You Only Pass Once: Answering and Abstaining Together in a Single Forward Pass of a Frozen Language Model

Measuring Fairness in Large Audio Language Models via Semantic-Aware Bias Estimation

Beyond Relevance-Centric Retrieval: Rubric-Oriented Document Set Selection and Ranking

OmegaUse-OfficeVal: Benchmarking LLM Agents on Long-Horizon Office-Suite Tasks with Economic Grounding