arXiv cs.AI - 2026-06-05

418 items collected.

1. How Far Did They Go? The Persuasive Tactics of Covert LLM Agents in a Discontinued Field Experiment

Author: Kokil Jaidka, Saifuddin Ahmed
Published: 6/5/2026, 4:00:00 AM
Categories: cs.AI

arXiv:2606.05256v1 Announce Type: new Abstract: This study analyzes a publicly released dataset from a discontinued field experiment on Reddit's r/ChangeMyView. The intervention, conducted by unknown, external researchers and halted following ethical backlash, involved undisclosed AI-generated accou...

arXiv cs.AI - 2026-06-05 ​

1. How Far Did They Go? The Persuasive Tactics of Covert LLM Agents in a Discontinued Field Experiment ​

2. What Should Agents Say? Action-state Communication for Efficient Multi-Agent Systems ​

3. I Know What You Meme, Even If it Emerged Today: Understanding Evolving Memes through Open-World Knowledge Acquisition ​

4. GITCO: Gated Inference-Time Context Optimization in TSFMs ​

5. Uncertainty Aware Functional Behavior Prediction and Material Fatigue Assessment for Circular Factory ​

6. SentinelBench: A Benchmark for Long-Running Monitoring Agents ​

7. An interpretable and trustworthy AI framework for large-scale longitudinal structure-pain association studies using data from the Osteoarthritis Initiative (OAI) ​

8. Synthetic Contrastive Reasoning for Multi-Table Q&A ​

9. Stability vs. Manipulability: Evaluating Robustness Under Post-Decision Interaction in LLM Judges ​

10. Residual Modeling for High-Fidelity Learned Compression of Scientific Data ​

11. LeanMarathon: Toward Reliable AI Co-Mathematicians through Long-Horizon Lean Autoformalization ​

12. Harnessing Generalist Agents for Contextualized Time Series ​

13. Agents' Last Exam ​

14. Mutation Without Variation: Convergence Dynamics in LLM-Driven Program Evolution ​

15. A Motivational Architecture for Conversational AGI ​

16. Assessing the Carbon Emissions and Energy Consumption of U.S. Hyperscale Data Centers ​

17. Minimizing the Hidden Cost of Scales: Graph-Guided Ultra-Low-Bit Quantization for Large Language Models ​

18. Zero knowledge verification for frontier AI training is possible ​

19. Ten Headache Specialists versus Artificial Intelligence for Clinical Literature Summarization: A Critical Evaluation and Comparison ​

20. Brick-Composer: Using MLLMs for Assembly with Diverse Bricks ​

21. Insurance of Agentic AI ​

22. Output Type Before Quality: A Standards-Derived XAI Admissibility Rubric for Autonomous-Driving Safety ​

23. PSEBench: A Controllable and Verifiable Benchmark for Evaluating LLMs in Patient Safety Event Triage ​

24. Step-by-Step Optimization-like Reasoning in LLMs over Expanding Search Spaces ​

25. Severity-Aware Curriculum Learning with Multi-Model Response Selection for Medical Text Generation ​

26. EpiEvolve: Self-Evolving Agents for Streaming Pandemic Forecasting under Regime Shifts ​

27. SciVisAgentSkills: Design and Evaluation of Agent Skills for Scientific Data Analysis and Visualization ​

28. When Should We Protect AI? A Precautionary Framework for Consciousness Uncertainty ​

29. Individual Gain, Collective Loss: Metacognitive Adaptation in AI-Assisted Creativity ​

30. SoCRATES: Towards Reliable Automated Evaluation of Proactive LLM Mediation across Domains and Socio-cognitive Variations ​

31. GuardNet: Ensemble Strategies of Shallow Neural Networks for Robust Prompt Injection and Jailbreak Detection ​

32. Fix the Mind, Not the Move: Interpretable AI Assistance via Knowledge-Gap Localization ​

33. Multilingual Fine-Tuning via Localized Gradient Conflict Resolution ​

34. Safety Paradox: How Enhanced Safety Awareness Leaves LLMs Vulnerable to Posterior Attack ​

35. Self-Commitment Latency: A Reward-Free Probe for Prompted Implicit Hacking ​

36. Evaluation of LLMs for Mathematical Formalization in Lean ​

37. Answer Presence Drives RAG Rewriting Gains ​

38. FIDES: Faithful Inference via Deep Evidence Signals for Retrieval-Memory Conflict in RAG ​

39. Coding with "Enemy": Can Human Developers Detect AI Agent Sabotage? ​

40. Continual Learning Bench: Evaluating Frontier AI Systems in Real-World Stateful Environments ​

41. Do More Agents Help? Controlled and Protocol-Aligned Evaluation of LLM Agent Workflows ​

42. Beyond Output Matching: Preserving Internal Geometry in NVFP4 LLM Distillatio ​

43. AdaMEM: Test-Time Adaptive Memory for Language Agents ​

44. PerceptUI: LLM Agents as Human-Aligned Synthetic Users for UI/UX Evaluation ​

45. Seeing Time: Benchmarking Chronological Reasoning and Shortcut Biases in Vision-Language Models ​

46. Critic-Guided Heterogeneous Multi-Agent Reasoning for Reliable Mathematical Problem Solving ​

47. DiG-Plan: Mitigating Early Commitment for Tool-Graph Planning via Diffusion Guidance ​

48. When AI Says It Feels ​

49. Class-Specific Branch Attention for Mitigating Gradient Interference under Class Imbalance ​

50. SubtleMemory: A Benchmark for Fine-Grained Relational Memory Discrimination in Long-Horizon AI Agents ​

51. TAPO: Tool-Aware Policy Optimization via Credit Transfer for Multimodal Search Agents ​

52. Can LLMs Write Correct TLA+ Specifications? Evaluating Natural-Language-to-TLA+ Generation ​

53. From Risk Classification to Action Plan Remediation: A Guardrail Feedback Driven Framework for LLM Agents ​

54. When Tools Fail: Benchmarking Dynamic Replanning and Anomaly Recovery in LLM Agents ​

55. Statistical Priors for Implicit Preferences: Decoupling Skill Selection as a Local Harness in Personal Agents ​

56. Agentic Molecular Recovery via Molecule-Aware Exploration ​

57. Entropy-Based Evaluation of AI Agents: A Lightweight Framework for Measuring Behavioral Patterns ​

58. QCFuse: Query-Aware Cache Fusion via Compressed View for Efficient RAG Serving ​

59. Retry Policy Gradients in Continuous Action Spaces ​

60. Retrospective Harness Optimization: Improving LLM Agents via Self-Preference over Trajectory Rollouts ​

61. Towards World Models in Biomedical Research ​

62. A Pre-Registered Causal Partition of Self-Consistency Elicitation and Reward Design in RLVR ​

63. Edit-R2: Context-Aware Reinforcement Learning for Multi-Turn Image Editing ​

64. Bidirectional Search for Longest Paths: Case for Front-to-Front Heuristics ​

65. The Self-Correction Illusion: LLMs Correct Others but Not Themselves ​

66. Framing, Judging, Steering: An Assessable Competency Model for Teach-ing Students to Reason With Generative AI ​

67. Beyond Vector Similarity: A Structural Analysis of Graph-Augmented Retrieval for Industrial Knowledge Graphs ​

68. PLAN-S: Bridging Planning with Latent Style Dynamics for Autonomous Driving World Models ​

69. RedditPersona: A Modular Framework for Community-Conditioned LLM Adaptation from Reddit ​

70. Memory is Reconstructed, Not Retrieved: Graph Memory for LLM Agents ​

71. Beyond Similarity: Trustworthy Memory Search for Personal AI Agents ​

72. When Should Memory Stay Silent: Measuring Memory-Use Boundaries in Memory-Augmented Conversational Agents ​

73. Learning Visual Spatial Planning from Symbolic State via Modality-Gap-Aware Self-Distillation ​

74. A Framework for Measuring Appropriate Reliance on Set-Valued AI Advice ​

75. Beyond Semantic Organization: Memory as Execution State Management for Long-Horizon Agents ​

76. Integrating Mechanistic and Data-Driven Models for Neurological Disorders through Differentiable Programming ​

77. CogManip: Benchmarking Manipulative Behavior in Multi-Turn Interactions with Large Language Model ​

78. Step-adaptive multimodal fusion network with multi-scale cloud feature learning for ultra-short-term solar irradiance forecasting ​

79. Towards Healthy Evolution: Exploring the Role and Mechanisms of Human-Agent Interaction in Self-Evolving Systems ​

arXiv cs.AI - 2026-06-05

1. How Far Did They Go? The Persuasive Tactics of Covert LLM Agents in a Discontinued Field Experiment

2. What Should Agents Say? Action-state Communication for Efficient Multi-Agent Systems

3. I Know What You Meme, Even If it Emerged Today: Understanding Evolving Memes through Open-World Knowledge Acquisition

4. GITCO: Gated Inference-Time Context Optimization in TSFMs

5. Uncertainty Aware Functional Behavior Prediction and Material Fatigue Assessment for Circular Factory

6. SentinelBench: A Benchmark for Long-Running Monitoring Agents

7. An interpretable and trustworthy AI framework for large-scale longitudinal structure-pain association studies using data from the Osteoarthritis Initiative (OAI)

8. Synthetic Contrastive Reasoning for Multi-Table Q&A

9. Stability vs. Manipulability: Evaluating Robustness Under Post-Decision Interaction in LLM Judges

10. Residual Modeling for High-Fidelity Learned Compression of Scientific Data

11. LeanMarathon: Toward Reliable AI Co-Mathematicians through Long-Horizon Lean Autoformalization

12. Harnessing Generalist Agents for Contextualized Time Series

13. Agents' Last Exam

14. Mutation Without Variation: Convergence Dynamics in LLM-Driven Program Evolution

15. A Motivational Architecture for Conversational AGI

16. Assessing the Carbon Emissions and Energy Consumption of U.S. Hyperscale Data Centers

17. Minimizing the Hidden Cost of Scales: Graph-Guided Ultra-Low-Bit Quantization for Large Language Models

18. Zero knowledge verification for frontier AI training is possible

19. Ten Headache Specialists versus Artificial Intelligence for Clinical Literature Summarization: A Critical Evaluation and Comparison

20. Brick-Composer: Using MLLMs for Assembly with Diverse Bricks

21. Insurance of Agentic AI

22. Output Type Before Quality: A Standards-Derived XAI Admissibility Rubric for Autonomous-Driving Safety

23. PSEBench: A Controllable and Verifiable Benchmark for Evaluating LLMs in Patient Safety Event Triage

24. Step-by-Step Optimization-like Reasoning in LLMs over Expanding Search Spaces

25. Severity-Aware Curriculum Learning with Multi-Model Response Selection for Medical Text Generation

26. EpiEvolve: Self-Evolving Agents for Streaming Pandemic Forecasting under Regime Shifts

27. SciVisAgentSkills: Design and Evaluation of Agent Skills for Scientific Data Analysis and Visualization

28. When Should We Protect AI? A Precautionary Framework for Consciousness Uncertainty

29. Individual Gain, Collective Loss: Metacognitive Adaptation in AI-Assisted Creativity

30. SoCRATES: Towards Reliable Automated Evaluation of Proactive LLM Mediation across Domains and Socio-cognitive Variations

31. GuardNet: Ensemble Strategies of Shallow Neural Networks for Robust Prompt Injection and Jailbreak Detection

32. Fix the Mind, Not the Move: Interpretable AI Assistance via Knowledge-Gap Localization

33. Multilingual Fine-Tuning via Localized Gradient Conflict Resolution

34. Safety Paradox: How Enhanced Safety Awareness Leaves LLMs Vulnerable to Posterior Attack

35. Self-Commitment Latency: A Reward-Free Probe for Prompted Implicit Hacking

36. Evaluation of LLMs for Mathematical Formalization in Lean

37. Answer Presence Drives RAG Rewriting Gains

38. FIDES: Faithful Inference via Deep Evidence Signals for Retrieval-Memory Conflict in RAG

39. Coding with "Enemy": Can Human Developers Detect AI Agent Sabotage?

40. Continual Learning Bench: Evaluating Frontier AI Systems in Real-World Stateful Environments

41. Do More Agents Help? Controlled and Protocol-Aligned Evaluation of LLM Agent Workflows

42. Beyond Output Matching: Preserving Internal Geometry in NVFP4 LLM Distillatio

43. AdaMEM: Test-Time Adaptive Memory for Language Agents

44. PerceptUI: LLM Agents as Human-Aligned Synthetic Users for UI/UX Evaluation

45. Seeing Time: Benchmarking Chronological Reasoning and Shortcut Biases in Vision-Language Models

46. Critic-Guided Heterogeneous Multi-Agent Reasoning for Reliable Mathematical Problem Solving

47. DiG-Plan: Mitigating Early Commitment for Tool-Graph Planning via Diffusion Guidance

48. When AI Says It Feels

49. Class-Specific Branch Attention for Mitigating Gradient Interference under Class Imbalance

50. SubtleMemory: A Benchmark for Fine-Grained Relational Memory Discrimination in Long-Horizon AI Agents

51. TAPO: Tool-Aware Policy Optimization via Credit Transfer for Multimodal Search Agents

52. Can LLMs Write Correct TLA+ Specifications? Evaluating Natural-Language-to-TLA+ Generation

53. From Risk Classification to Action Plan Remediation: A Guardrail Feedback Driven Framework for LLM Agents

54. When Tools Fail: Benchmarking Dynamic Replanning and Anomaly Recovery in LLM Agents

55. Statistical Priors for Implicit Preferences: Decoupling Skill Selection as a Local Harness in Personal Agents

56. Agentic Molecular Recovery via Molecule-Aware Exploration

57. Entropy-Based Evaluation of AI Agents: A Lightweight Framework for Measuring Behavioral Patterns

58. QCFuse: Query-Aware Cache Fusion via Compressed View for Efficient RAG Serving

59. Retry Policy Gradients in Continuous Action Spaces

60. Retrospective Harness Optimization: Improving LLM Agents via Self-Preference over Trajectory Rollouts

61. Towards World Models in Biomedical Research

62. A Pre-Registered Causal Partition of Self-Consistency Elicitation and Reward Design in RLVR

63. Edit-R2: Context-Aware Reinforcement Learning for Multi-Turn Image Editing

64. Bidirectional Search for Longest Paths: Case for Front-to-Front Heuristics

65. The Self-Correction Illusion: LLMs Correct Others but Not Themselves

66. Framing, Judging, Steering: An Assessable Competency Model for Teach-ing Students to Reason With Generative AI

67. Beyond Vector Similarity: A Structural Analysis of Graph-Augmented Retrieval for Industrial Knowledge Graphs

68. PLAN-S: Bridging Planning with Latent Style Dynamics for Autonomous Driving World Models

69. RedditPersona: A Modular Framework for Community-Conditioned LLM Adaptation from Reddit

70. Memory is Reconstructed, Not Retrieved: Graph Memory for LLM Agents

71. Beyond Similarity: Trustworthy Memory Search for Personal AI Agents

72. When Should Memory Stay Silent: Measuring Memory-Use Boundaries in Memory-Augmented Conversational Agents

73. Learning Visual Spatial Planning from Symbolic State via Modality-Gap-Aware Self-Distillation

74. A Framework for Measuring Appropriate Reliance on Set-Valued AI Advice

75. Beyond Semantic Organization: Memory as Execution State Management for Long-Horizon Agents

76. Integrating Mechanistic and Data-Driven Models for Neurological Disorders through Differentiable Programming

77. CogManip: Benchmarking Manipulative Behavior in Multi-Turn Interactions with Large Language Model

78. Step-adaptive multimodal fusion network with multi-scale cloud feature learning for ultra-short-term solar irradiance forecasting

79. Towards Healthy Evolution: Exploring the Role and Mechanisms of Human-Agent Interaction in Self-Evolving Systems