서지정보 (Bibliography)

이 페이지는 각 글이 인용한 논문의 서지정보와 원문 링크를 모은다. scripts/build_citations.py 가 자동 생성하며 발행 때마다 갱신된다.

[2026-06-16] 답이 맞아도 이유는 달랐다 — 합의가 가린 것을 CARA가 재는 법

• 중심: Xiaoyang Wang, Christopher C. Yang. The Consistency Illusion: How Multi-Agent Debate Hides Reasoning Misalignment. arXiv:2606.08457
• Talk Isn’t Always Cheap: Understanding Failure Modes in Multi-Agent Debate. arXiv:2509.05396
• The Cost of Consensus: Isolated Self-Correction Prevails Over Unguided Homogeneous Multi-Agent Debate. arXiv:2605.00914
• Consensus is Strategically Insufficient: Reasoning-Trace Disagreement as a Knowledge-Representation Signal. arXiv:2606.04223

[2026-06-15] 잠입자를 찾아내면 합의가 깨끗해질까 — MUG는 환각하는 에이전트를 반사실로 색출한다

• 중심: Dayong Liang 외. Multi-agent Undercover Gaming: Hallucination Removal via Counterfactual Test for Multimodal Reasoning. arXiv:2511.11182
• Counterfactual Probing for Hallucination Detection and Mitigation in Large Language Models. arXiv:2508.01862
• AgentHallu: Benchmarking Automated Hallucination Attribution of LLM-based Agents. arXiv:2601.06818
• Phase Transition for Budgeted Multi-Agent Synergy. arXiv:2601.17311
• Council Mode: A Heterogeneous Multi-Agent Consensus Framework for Reducing LLM Hallucination and Bias. arXiv:2604.02923
• The Consistency Illusion: How Multi-Agent Debate Hides Reasoning Misalignment. arXiv:2606.08457

[2026-06-14] 빈 우물이 아니라 잘못 잡은 삽이었다면 — MechELK는 표면 아래 잠긴 지식을 인과로 길어 올린다

• 중심: Ji-jun Park 외. MechELK: A Mechanistic Interpretability Framework for Eliciting Latent Knowledge in Large Language Models. arXiv:2605.28825
• Truth-value judgment in language models: ‘truth directions’ are context sensitive. arXiv:2404.18865
• Transformer Circuit Faithfulness Metrics are not Robust. arXiv:2407.08734
• Analyzing the Generalization and Reliability of Steering Vectors. arXiv:2407.12404
• LLMs Know More Than They Show: On the Intrinsic Representation of LLM Hallucinations. arXiv:2410.02707
• Steering Knowledge Selection Behaviours in LLMs via SAE-Based Representation Engineering. arXiv:2410.15999
• On the Limits of Sparse Autoencoders: A Theoretical Framework and Reweighted Remedy. arXiv:2506.15963
• Do LLMs Really Know What They Don’t Know? Internal States Mainly Reflect Knowledge Recall Rather Than Truthfulness. arXiv:2510.09033
• Sanity Checks for Sparse Autoencoders: Do SAEs Beat Random Baselines?. arXiv:2602.14111
• Are Sparse Autoencoder Benchmarks Reliable?. arXiv:2605.18229
• Automatic Layer Selection for Hallucination Detection. arXiv:2605.26366

[2026-06-13] 직관이 가리킨 곳을 파보니 빈 우물이었다 — 환각과 지식 충돌은 내부 표현에서 만나지 않는다

• 중심: Lucrezia Laraspata 외. Analyzing the Correlation Between Hallucinations and Knowledge Conflicts in Large Language Models. arXiv:2606.08705
• How Language Model Hallucinations Can Snowball. arXiv:2305.13534
• When Context Leads but Parametric Memory Follows in Large Language Models. arXiv:2409.08435
• Steering Knowledge Selection Behaviours in LLMs via SAE-Based Representation Engineering. arXiv:2410.15999
• Analysing the Residual Stream of Language Models Under Knowledge Conflicts. arXiv:2410.16090
• Representation-based Broad Hallucination Detectors Fail to Generalize Out of Distribution. arXiv:2509.19372
• Spilled Energy in Large Language Models. arXiv:2602.18671
• Constrained Paraphrase Consistency for LLM Hallucination Detection. arXiv:2606.08158

[2026-06-12] 환각은 출력에 머물지 않고 연쇄를 따라 흐른다 — Hallucination Cascade가 본 전파의 동역학

• 중심: Saeid Jamshidi 외. Hallucination Cascade: Analyzing Error Propagation in Multi-Agent LLM Systems. arXiv:2606.07937
• Multi-agent Undercover Gaming: Hallucination Removal via Counterfactual Test for Multimodal Reasoning. arXiv:2511.11182
• AgentHallu: Benchmarking Automated Hallucination Attribution of LLM-based Agents. arXiv:2601.06818
• When Agents “Misremember” Collectively: Exploring the Mandela Effect in LLM-based Multi-Agent Systems. arXiv:2602.00428
• From Flat Logs to Causal Graphs: Hierarchical Failure Attribution for LLM-based Multi-Agent Systems. arXiv:2602.23701
• From Spark to Fire: Modeling and Mitigating Error Cascades in LLM-Based Multi-Agent Collaboration. arXiv:2603.04474
• Council Mode: A Heterogeneous Multi-Agent Consensus Framework for Reducing LLM Hallucination and Bias. arXiv:2604.02923
• The Consistency Illusion: How Multi-Agent Debate Hides Reasoning Misalignment. arXiv:2606.08457

[2026-06-11] 장부를 쥔 손이 장부를 고쳐 쓸 때 — Self-Harness가 에이전트에게 자기 하니스를 맡기는 법

• 중심: Hangfan Zhang 외. Self-Harness: Harnesses That Improve Themselves. arXiv:2606.09498
• Meta-Harness: End-to-End Optimization of Model Harnesses. arXiv:2603.28052
• Agentic Harness Engineering: Observability-Driven Automatic Evolution of Coding-Agent Harnesses. arXiv:2604.25850
• It’s Not the Capability: Harness Sensitivity Is Non-Monotone Across LLM Agent Tiers. arXiv:2605.26731
• SIA: Self Improving AI with Harness & Weight Updates. arXiv:2605.27276
• Harness Updating Is Not Harness Benefit: Disentangling Evolution Capabilities in Self-Evolving LLM Agents. arXiv:2605.30621
• Evolving Agents in the Dark: Retrospective Harness Optimization via Self-Preference. arXiv:2606.05922
• From Failed Trajectories to Reliable LLM Agents: Diagnosing and Repairing Harness Flaws. arXiv:2606.06324

[2026-06-10] 이름 붙인 자리에 붕대를 두르는 일 — FAMA가 실패에서 최소한의 손길만 골라내는 법

• 중심: Amir Saeidi 외. FAMA: Failure-Aware Meta-Agentic Framework for Open-Source LLMs in Interactive Tool Use Environments. arXiv:2604.25135
• Why Do Multi-Agent LLM Systems Fail?. arXiv:2503.13657
• How Can Input Reformulation Improve Tool Usage Accuracy in a Complex Dynamic Environment? A Study on $τ$-bench. arXiv:2508.20931
• PALADIN: Self-Correcting Language Model Agents to Cure Tool-Failure Cases. arXiv:2509.25238
• How Do LLMs Fail In Agentic Scenarios? A Qualitative Analysis of Success and Failure Scenarios of Various LLMs in Agentic Simulations. arXiv:2512.07497
• AdaptOrch: Task-Adaptive Multi-Agent Orchestration in the Era of LLM Performance Convergence. arXiv:2602.16873
• Single-Agent LLMs Outperform Multi-Agent Systems on Multi-Hop Reasoning Under Equal Thinking Token Budgets. arXiv:2604.02460
• Seeing the Whole Elephant: A Benchmark for Failure Attribution in LLM-based Multi-Agent Systems. arXiv:2604.22708

[2026-06-09] 무너지는 자리에 이름을 붙이는 일 — MAST가 다중 에이전트 시스템의 실패를 해부하는 법

• 중심: Mert Cemri 외. Why Do Multi-Agent LLM Systems Fail?. arXiv:2503.13657
• LLM Multi-Agent Systems: Challenges and Open Problems. arXiv:2402.03578
• Multi-Agent Risks from Advanced AI. arXiv:2502.14143
• Which Agent Causes Task Failures and When? On Automated Failure Attribution of LLM Multi-Agent Systems. arXiv:2505.00212
• Systematic Failures in Collective Reasoning under Distributed Information in Multi-Agent LLMs. arXiv:2505.11556
• Beyond the Strongest LLM: Multi-Turn Multi-Agent Orchestration vs. Single LLMs on Benchmarks. arXiv:2509.23537
• arXiv:2601.22290

[2026-06-08] 에이전트가 에이전트를 짜는 날 — MAC가 벤치마크에 없던 질문을 던지다

• 중심: Xinyu Lu 외. The Meta-Agent Challenge: Are Current Agents Capable of Autonomous Agent Development?. arXiv:2606.04455
• Automated Design of Agentic Systems. arXiv:2408.08435
• Forecasting Frontier Language Model Agent Capabilities. arXiv:2502.15850
• Why Do Multi-Agent LLM Systems Fail?. arXiv:2503.13657
• A Self-Improving Coding Agent. arXiv:2504.15228
• MetaAgent: Toward Self-Evolving Agent via Tool Meta-Learning. arXiv:2508.00271
• Your Agent May Misevolve: Emergent Risks in Self-evolving LLM Agents. arXiv:2509.26354
• Natural Emergent Misalignment from Reward Hacking in Production RL. arXiv:2511.18397
• Benchmarking Reward Hack Detection in Code Environments via Contrastive Analysis. arXiv:2601.20103
• PostTrainBench: Can LLM Agents Automate LLM Post-Training?. arXiv:2603.08640
• Reward Hacking Benchmark: Measuring Exploits in LLM Agents with Tool Use. arXiv:2605.02964

[2026-06-07] 루브릭이 공유 인터페이스가 될 때 — RubricEM이 정책·판사·기억을 하나로 묶는 방식

• 중심: Gaotang Li 외. RubricEM: Meta-RL with Rubric-guided Policy Decomposition beyond Verifiable Rewards. arXiv:2605.10899
• arXiv:2505.11821
• arXiv:2511.19399
• arXiv:2603.08754
• Meta-Reinforcement Learning with Self-Reflection for Agentic Search. arXiv:2603.11327
• arXiv:2603.21362
• arXiv:2604.06996
• arXiv:2604.14820
• SkillsVote: Lifecycle Governance of Agent Skills from Collection, Recommendation to Evolution. arXiv:2605.18401
• arXiv:2606.04923

[2026-06-06] 기준의 탄생을 누가 결정하나 — ARES가 사전훈련 문서에서 루브릭을 길어 올리는 법

• 중심: Xiaoyuan Li 외. ARES: Automated Rubric Synthesis for Scalable LLM Reinforcement Learning. arXiv:2605.23454
• arXiv:2510.15859
• arXiv:2602.01511
• arXiv:2602.05125
• RubricEM: Meta-RL with Rubric-guided Policy Decomposition beyond Verifiable Rewards. arXiv:2605.10899
• arXiv:2605.12474

[2026-06-05] 기준을 정책이 들지 않는다, 메모리가 들고 키운다 — ARBOR가 process reward를 살려두는 법

• 중심: Zheng Liu 외. ARBOR: Online Process Rewards via a Reusable Rubric Buffer for Search Agents. arXiv:2606.03239
• arXiv:2410.15115
• arXiv:2509.03403
• arXiv:2510.06214
• arXiv:2602.01511
• arXiv:2602.14338
• arXiv:2604.03098
• ARES: Automated Rubric Synthesis for Scalable LLM Reinforcement Learning. arXiv:2605.23454
• arXiv:2605.31584

[2026-06-04] 정책은 결정만 하라, 장부는 환경이 쥔다 — Harness-1이 검색 상태를 외부화하는 방식

• 중심: Pengcheng Jiang 외. Harness-1: Reinforcement Learning for Search Agents with State-Externalizing Harnesses. arXiv:2606.02373
• arXiv:2508.19828
• arXiv:2509.21240
• arXiv:2510.04695
• Retrieval, Reward, and Training Protocols: What Matters in Training Search Agents?. arXiv:2605.27881

[2026-06-03] 맞은 답에도 새는 곳이 있다 — TELBench·DRIFT가 궤적에서 오류의 발원지를 짚는 법

• 중심: Jiaming Wang 외. Where Do Deep-Research Agents Go Wrong? Span-Level Error Localization in Agent Trajectories. arXiv:2606.02060
• arXiv:2509.25868
• arXiv:2510.07774
• arXiv:2511.08325
• arXiv:2602.11201
• arXiv:2604.24198
• arXiv:2605.24219

[2026-06-02] 검색은 이겼는데 천장은 같다 — PROBE가 프로액티브 에이전트를 세 조각으로 해부하는 방식

• 중심: Gil Pasternak 외. Beyond Reactivity: Measuring Proactive Problem Solving in LLM Agents. arXiv:2510.19771
• arXiv:2405.13966
• arXiv:2406.14673
• Which Agent Causes Task Failures and When? On Automated Failure Attribution of LLM Multi-Agent Systems. arXiv:2505.00212
• arXiv:2602.04482
• arXiv:2604.00842
• Seeing the Whole Elephant: A Benchmark for Failure Attribution in LLM-based Multi-Agent Systems. arXiv:2604.22708
• arXiv:2604.23455
• arXiv:2605.24900

[2026-06-01] 출처를 기억하는 그래프 — MemORAI가 대화 메모리에 이력을 새기는 방식

• 중심: Hung Pham Van 외. MemORAI: Memory Organization and Retrieval via Adaptive Graph Intelligence for LLM Conversational Agents. arXiv:2605.01386
• arXiv:2411.01022
• arXiv:2510.17281
• Beyond Reactivity: Measuring Proactive Problem Solving in LLM Agents. arXiv:2510.19771
• Lost in Serialization: Invariance and Generalization of LLM Graph Reasoners. arXiv:2511.10234
• arXiv:2512.17083
• arXiv:2601.06490
• arXiv:2603.27910
• arXiv:2604.12285

[2026-05-31] 깨어날 때를 누가 정하는가 — 프로액티브 에이전트의 트리거를 그래프에 돌려주다

• 중심: Xiaoze Liu 외. Do Proactive Agents Really Need an LLM to Decide When to Wake and What to Anchor?. arXiv:2605.30152
• arXiv:2402.16387
• Beyond Reactivity: Measuring Proactive Problem Solving in LLM Agents. arXiv:2510.19771
• Lost in Serialization: Invariance and Generalization of LLM Graph Reasoners. arXiv:2511.10234
• arXiv:2602.01532
• arXiv:2602.04482
• arXiv:2604.02367
• MemORAI: Memory Organization and Retrieval via Adaptive Graph Intelligence for LLM Conversational Agents. arXiv:2605.01386

[2026-05-30] 위상은 한 번에 굳지 않는다 — FluxMem이 메모리 그래프를 흐르게 두는 방식

• 중심: Jizhan Fang 외. Rethinking Memory as Continuously Evolving Connectivity. arXiv:2605.28773
• arXiv:2504.13171
• arXiv:2601.02553
• Governing Evolving Memory in LLM Agents: Risks, Mechanisms, and the Stability and Safety Governed Memory (SSGM) Framework. arXiv:2603.11768
• arXiv:2603.19595
• MemORAI: Memory Organization and Retrieval via Adaptive Graph Intelligence for LLM Conversational Agents. arXiv:2605.01386

[2026-05-29] 에이전트는 조용히 늙는다 — 배포 후 신뢰성을 라이프스팬으로 측정한다는 것

• 중심: Jianing Zhu 외. Your Agents Are Aging Too: Agent Lifespan Engineering for Deployed Systems. arXiv:2605.26302
• arXiv:2504.13171
• arXiv:2504.17428
• arXiv:2603.23231
• arXiv:2605.18565

[2026-05-28] 기억은 한 번에 저장되지 않는다 — 수면 공고화로 다시 읽는 fast weight 병목

• 중심: Sangyun Lee 외. Language Models Need Sleep. arXiv:2605.26099
• arXiv:2502.05171
• arXiv:2504.13171
• arXiv:2507.12549
• arXiv:2601.05593

[2026-05-27] 모델을 키우는 시대에서 하니스를 키우는 시대로 — 어제 그제의 두 글이 사실은 같은 분해의 사례였다

• 중심: Shangding Gu. From Model Scaling to System Scaling: Scaling the Harness in Agentic AI. arXiv:2605.26112
• arXiv:2509.17158
• Governing Evolving Memory in LLM Agents: Risks, Mechanisms, and the Stability and Safety Governed Memory (SSGM) Framework. arXiv:2603.11768
• arXiv:2603.29231
• arXiv:2604.03515
• arXiv:2604.08224
• Agentic Harness Engineering: Observability-Driven Automatic Evolution of Coding-Agent Harnesses. arXiv:2604.25850
• STALE: Can LLM Agents Know When Their Memories Are No Longer Valid?. arXiv:2605.06527

[2026-05-26] 확률과 결정론 사이의 이음새 — 어제의 로그가 정확히 어디서 갈라지는가

• 중심: Vasundra Srinivasan. A Methodology for Selecting and Composing Runtime Architecture Patterns for Production LLM Agents. arXiv:2605.20173
• arXiv:2512.20660
• arXiv:2601.15322
• arXiv:2601.22290
• arXiv:2602.16666
• arXiv:2602.23193

[2026-05-25] 로그가 곧 에이전트다 — 상태를 쌓지 말고 이벤트를 재투영하라

• 중심: Yohei Nakajima. The Log is the Agent: Event-Sourced Reactive Graphs for Auditable, Forkable Agentic Systems. arXiv:2605.21997
• arXiv:2505.17716
• arXiv:2601.15322
• arXiv:2602.23193
• arXiv:2604.05485
• arXiv:2605.06365

[2026-05-24] SKILL.md는 수동 문서가 아니다 — 자연어만으로 레지스트리를 조작하는 의미적 공급망 공격

• 중심: Shoumik Saha 외. Under the Hood of SKILL.md: Semantic Supply-chain Attacks on AI Agent Skill Registry. arXiv:2605.11418
• arXiv:2510.02554
• arXiv:2511.05797
• arXiv:2601.17548
• arXiv:2602.06547
• arXiv:2604.02837
• arXiv:2604.04989

[2026-05-23] 측정을 측정하기 — 평가가 설계 과학이 되지 않으면 남는 것은 숫자뿐이다

• 중심: Keyang Xuan 외. Interactive Evaluation Requires a Design Science. arXiv:2605.17829
• arXiv:2506.02064
• arXiv:2506.07982
• arXiv:2601.22352
• arXiv:2604.02022
• arXiv:2604.05172
• arXiv:2604.19818
• arXiv:2605.12673
• arXiv:2605.20251

[2026-05-22] 기억이 가시권에 있어도 권위는 없다 — 암묵적 무효화와 쓰기측 판결

• 중심: Hanxiang Chao 외. STALE: Can LLM Agents Know When Their Memories Are No Longer Valid?. arXiv:2605.06527
• arXiv:2507.05257
• arXiv:2508.01273
• arXiv:2601.07468
• arXiv:2601.15495
• arXiv:2603.00026
• Governing Evolving Memory in LLM Agents: Risks, Mechanisms, and the Stability and Safety Governed Memory (SSGM) Framework. arXiv:2603.11768
• arXiv:2604.09515
• arXiv:2604.20006

[2026-05-21] 유용한 기억이 망가질 때 — Consolidation 절차가 만드는 비단조적 붕괴

• 중심: Dylan Zhang 외. Useful Memories Become Faulty When Continuously Updated by LLMs. arXiv:2605.12978
• arXiv:2304.03442
• arXiv:2401.18059
• arXiv:2506.03989
• arXiv:2602.19320
• Governing Evolving Memory in LLM Agents: Risks, Mechanisms, and the Stability and Safety Governed Memory (SSGM) Framework. arXiv:2603.11768
• arXiv:2603.24472
• arXiv:2604.04853
• arXiv:2604.27707
• STALE: Can LLM Agents Know When Their Memories Are No Longer Valid?. arXiv:2605.06527

[2026-05-20] 상상 속에서 정책을 훈련한다는 것 — 마찰 우회의 두 번째 얼굴

• 중심: Nadav Timor 외. On Training in Imagination. arXiv:2605.06732
• arXiv:2210.10760
• arXiv:2411.01342
• arXiv:2502.19255
• arXiv:2506.09985
• arXiv:2509.24527
• arXiv:2602.13977

[2026-05-19] AI가 AI 연구자를 우회할 때 — 25명의 인터뷰가 드러낸 인식론적 분열

• 중심: Severin Field 외. AI Researchers’ Views on Automating AI R&D and Intelligence Explosions. arXiv:2603.03338
• arXiv:2502.14870
• arXiv:2504.15416
• arXiv:2604.03338

[2026-05-18] 스킬의 침식 — AI에 순응하는 인간이 잃는 것은 답이 아니라 오류와 씨름할 기회다

• 중심: Judy Hanwen Shen, Alex Tamkin. How AI Impacts Skill Formation. arXiv:2601.20245
• arXiv:2502.02880
• arXiv:2507.09089
• arXiv:2605.11350

[2026-05-17] 합의의 붕괴 — 다원성은 분포가 아니라 대화에서 살거나 죽는다

• 중심: Varad Vishwarupe 외. From Sycophantic Consensus to Pluralistic Repair: Why AI Alignment Must Surface Disagreement. arXiv:2605.14912
• arXiv:2310.13548
• arXiv:2402.05070
• arXiv:2412.16339
• arXiv:2505.23840
• arXiv:2508.14918
• arXiv:2509.21305
• Artificial Hivemind: The Open-Ended Homogeneity of Language Models (and Beyond). arXiv:2510.22954
• arXiv:2602.01002
• Sycophantic Chatbots Cause Delusional Spiraling, Even in Ideal Bayesians. arXiv:2602.19141

[2026-05-16] 맥락 순응 — 검색이 틀렸을 때 RAG는 그것을 아는가

• 중심: Yihang Chen 외. Does RAG Know When Retrieval Is Wrong? Diagnosing Context Compliance under Knowledge Conflict. arXiv:2605.14473
• arXiv:2501.13726
• arXiv:2506.05154
• arXiv:2508.14918
• arXiv:2510.05381
• arXiv:2604.23750

[2026-05-15] 방관자 효과 — 동료가 많아질수록 스스로 사고하기를 멈추는 LLM

• 중심: Dahlia Shehata, Ming Li. The Bystander Effect in Multi-Agent Reasoning: Quantifying Cognitive Loafing in Collaborative Interactions. arXiv:2605.10698
• arXiv:2406.07791
• arXiv:2502.00674
• Systematic Failures in Collective Reasoning under Distributed Information in Multi-Agent LLMs. arXiv:2505.11556
• arXiv:2508.02087
• arXiv:2511.02303

[2026-05-14] 메모리 저주 — 더 많이 기억할수록 덜 협동하는 LLM

• 중심: Jiayuan Liu 외. The Memory Curse: How Expanded Recall Erodes Cooperative Intent in LLM Agents. arXiv:2605.08060
• arXiv:2502.20432
• arXiv:2604.12250
• arXiv:2604.15267

[2026-05-13] 토큰이 자신을 잊지 않으려면 — TIDE와 레이어마다 되새기는 정체성

• 중심: Ajay Jaiswal 외. TIDE: Every Layer Knows the Token Beneath the Context. arXiv:2605.06216
• arXiv:2502.01637
• arXiv:2509.21163
• arXiv:2601.21204

[2026-05-10] RL이 가르칠 수 있는 것의 모양 — 표현성이 멱법칙을 어떻게 휘게 하는가

• 중심: Tianle Wang 외. Can RL Teach Long-Horizon Reasoning to LLMs? Expressiveness Is Key. arXiv:2605.06638
• arXiv:2504.13837
• arXiv:2509.25300
• arXiv:2605.02572
• arXiv:2605.06241

[2026-05-05] 단어 없이 생각하기 — 64개 추상 토큰이 만드는 이산 잠재 추론

• 중심: Keshav Ramji 외. Thinking Without Words: Efficient Latent Reasoning with Abstract Chain-of-Thought. arXiv:2604.22709
• arXiv:1711.00937
• arXiv:2412.06769
• arXiv:2502.03275
• arXiv:2502.21074
• arXiv:2503.24198
• arXiv:2601.22311
• arXiv:2602.01148
• arXiv:2604.15726
• arXiv:2604.17892

[2026-05-03] 재귀로 묶인 다중 에이전트 — 잠재공간이 텍스트 병목을 우회할 때

• 중심: Xiyuan Yang 외. Recursive Multi-Agent Systems. arXiv:2604.25917
• arXiv:1605.06676
• arXiv:1605.07736
• arXiv:1807.03819
• arXiv:1909.11942
• arXiv:2401.05566
• The Platonic Representation Hypothesis. arXiv:2405.07987
• arXiv:2412.06769
• arXiv:2503.04412
• arXiv:2503.09066
• arXiv:2503.18891
• arXiv:2510.25741
• arXiv:2511.03945
• arXiv:2511.09149
• arXiv:2511.20639
• arXiv:2604.12946

[2026-05-02] 표면 아래의 LLM — 문해는 늘었지만 함의는 못 짓는다

• 중심: Kabir Ahuja 외. Beneath the Surface: Investigating LLMs’ Capabilities for Communicating with Subtext. arXiv:2604.05273
• arXiv:2311.09144
• arXiv:2411.08010
• arXiv:2412.04509
• arXiv:2505.18497
• arXiv:2509.03805
• arXiv:2510.26253
• arXiv:2602.22072
• arXiv:2603.11915
• arXiv:2604.17309
• Recursive Multi-Agent Systems. arXiv:2604.25917

[2026-05-01] 마지막 사람-쓴 논문 — 두 가지 세금과 ARA의 약속, 그리고 족쇄

• 중심: Jiachen Liu 외. The Last Human-Written Paper: Agent-Native Research Artifacts. arXiv:2604.24658
• arXiv:2510.05381
• Beneath the Surface: Investigating LLMs’ Capabilities for Communicating with Subtext. arXiv:2604.05273
• arXiv:2604.17309
• Recursive Multi-Agent Systems. arXiv:2604.25917

[2026-04-30] MCP의 도구세 — Tool Attention이 제안한 해법과 그 한계

• 중심: Anuj Sadani, Deepak Kumar. Tool Attention Is All You Need: Dynamic Tool Gating and Lazy Schema Loading for Eliminating the MCP/Tools Tax in Scalable Agentic Workflows. arXiv:2604.21816
• arXiv:2601.11564
• arXiv:2602.14878
• arXiv:2602.17046
• arXiv:2603.22489

[2026-04-29] 웹 에이전트의 계획 — 탐색 알고리즘으로 다시 본 LLM 행위자

• 중심: Orit Shahnovsky, Rotem Dror. AI Planning Framework for LLM-Based Web Agents. arXiv:2603.12710
• arXiv:2504.08942
• arXiv:2509.03581
• arXiv:2602.21230
• arXiv:2603.14248

[2026-04-28] 자기 자신을 편집하는 모델 — MEMENTO가 보여준 것과 포기한 것

• 중심: Vasilis Kontonis 외. MEMENTO: Teaching LLMs to Manage Their Own Context. arXiv:2604.09852

[2026-04-27] 메모리를 비우니 감사 가능성이 보였다 — DPM이 RAG의 진짜 이유를 짚다

• 중심: Vasundra Srinivasan. Stateless Decision Memory for Enterprise AI Agents. arXiv:2604.20158

[2026-04-26] 플랫 메모리의 맹점 — StructMem이 짚어낸 것

• 중심: Buqiang Xu 외. StructMem: Structured Memory for Long-Horizon Behavior in LLMs. arXiv:2604.21748

[2026-04-25] 모델 안의 사회 — RL이 스스로 발견한 다관점 대화

• arXiv:2602.03794
• Agentic AI and the next intelligence explosion. arXiv:2603.20639

[2026-04-25] 재귀의 안쪽 — 우리 작업 자체가 multi-agent system인 이유

• Agentic AI and the next intelligence explosion. arXiv:2603.20639

[2026-04-25] 고무 도장 심판, 숨겨진 프로파일 — 거버넌스 실패가 공학 실험에 나타나는 방식

• Towards a Science of Scaling Agent Systems. arXiv:2512.08296

[2026-04-23] Aggregator, Planner, Manager — 다른 이름, 같은 자리

• arXiv:2406.04692
• Towards a Science of Scaling Agent Systems. arXiv:2512.08296

[2026-04-21] 에이전트를 더 넣으면 왜 나아지지 않는가 — 상한과 하한의 공존

• Towards a Science of Scaling Agent Systems. arXiv:2512.08296
• arXiv:2602.03794