Search: “llm”

MemMorph: Tool Hijacking in LLM Agents via Memory Poisoning

arXiv:2605.26154v1 Announce Type: cross Abstract: LLM-driven agents are capable of selecting external tools to complete users' tasks. However, attackers could compromise such process, steering agents

Multi-Stakeholder LLM Alignment: Decomposing Estimation from Aggregation

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

DGX agent

arXiv:2605.26878v1 Announce Type: new Abstract: Multi-stakeholder tasks require one output to satisfy users with conflicting preferences. Holistic LLM judges conflate utility estimation and utility ag

safetyarxiv-cs-ai

27 May 2026

Shopping Companion: A Memory-Augmented LLM Agent for Real-World E-Commerce Tasks

arXiv:2603.14864v2 Announce Type: replace Abstract: In e-commerce, LLM agents show promise for shopping tasks such as recommendations, budget management, and bundle deals, where accurately capturing u

The MiniMax M2 series was one of the most widely used open-weight LLM series earlier this year. Now, we got a technical report with some int…

model-releasessebastian-raschka--x

The MiniMax M2 series was one of the most widely used open-weight LLM series earlier this year. Now, we got a technical report with some interesting tidbits. I summarized some of them below: 1. Full a

Vectors Are Not Neutral: Sensitive-Information Inference from Exported LLM Representations in Summarization

arXiv:2605.26433v1 Announce Type: new Abstract: Large language model (LLM) summarization systems may pass compact vector representations of private inputs to downstream retrieval, monitoring, audit, o

Agent-Facing Information Design in LLM Tool Registries

arXiv:2605.23916v1 Announce Type: cross Abstract: LLM tool registries function as unregulated advertising platforms: providers write free-text descriptions that agents use for selection, yet no measur

CausalFlow: Causal Attribution and Counterfactual Repair for LLM Agent Failures

arXiv:2605.25338v1 Announce Type: cross Abstract: Large language model (LLM) agents frequently fail on multi-step tasks involving reasoning, tool use, and environment interaction. While such failures

Local Ai

CONF-KV: Confidence-Aware KV Cache Eviction with Mixed-Precision Storage for Long-Horizon LLM

arXiv:2605.24786v1 Announce Type: cross Abstract: Long-horizon LLM inference turns the key--value (KV) cache into the dominant GPU memory consumer and makes per-token attention increasingly expensive.

local-aiarxiv-cs-ai

Empirical Analysis and Detection of Hallucinations in LLM-Generated Bug Report Summaries

arXiv:2605.24137v1 Announce Type: cross Abstract: Large Language Models (LLMs) are increasingly used to generate summaries of software bug reports, including sections such as Steps-to-Reproduce (S2R),

EVA: Accelerating LLM Decoding via an Efficient Vector Quantization Architecture

arXiv:2605.24144v1 Announce Type: cross Abstract: Large Language Models (LLMs) have achieved impressive performance across diverse domains but remain inefficient during the autoregressive decoding pha

researcharxiv-cs-lg

Evo-Attacker: Memory-Augmented Reinforcement Learning for Long-Horizon Tool Attacks on LLM-MAS

arXiv:2605.25389v1 Announce Type: cross Abstract: While Large Language Model-based Multi-Agent Systems (LLM-MAS) demonstrate remarkable capabilities in solving complex tasks by orchestrating specializ

Guarded Repair for Harm-Aware Post-hoc Replacement of LLM Mathematical Reasoning

arXiv:2605.24613v1 Announce Type: cross Abstract: Post-hoc repair of LLM mathematical reasoning introduces an asymmetric risk: fixing an incorrect reasoning trace is useful, but replacing a trace that

Influence-Inspired Spectral Rotations for Extreme Low-Bit LLM Quantization

arXiv:2605.25203v1 Announce Type: cross Abstract: We apply the influence-adaptive Walsh geometry of a companion theory paper (arXiv:2605.01637) to extreme low-bit weight-only LLM quantization. The rec

Local Ai

LLM-AutoSciLab: Closed-Loop Scientific Discovery via Active Experimentation with LLMs

arXiv:2605.24043v1 Announce Type: cross Abstract: Scientific discovery is a closed-loop process in which hypotheses guide data acquisition and observations refine the hypothesis space. Yet most approa

local-aiarxiv-cs-ai

Market Regime Council for Dynamic Credit Assignment in Multi-Agent LLM Decision Systems

arXiv:2605.24490v1 Announce Type: new Abstract: Multi-agent LLM decision systems for portfolio management still lack a principled way to assign credit across specialist agents, remain vulnerable to co

Measuring the Depth of LLM Unlearning via Activation Patching

arXiv:2605.24614v1 Announce Type: cross Abstract: Large language model (LLM) unlearning has emerged as a crucial post-hoc mechanism for privacy protection and AI safety, yet auditing whether target kn

Memory-Induced Tool-Drift in LLM Agents

arXiv:2605.24941v1 Announce Type: cross Abstract: Modern LLM agents combine long-term memory for personalization with tool-calling interfaces for taking actions in the world -- a combination underpinn

Mixture of Complementary Agents for Robust LLM Ensemble

arXiv:2605.24048v1 Announce Type: cross Abstract: Multi-AI collaboration, such as ensembling or debating large language models (LLMs), is a promising paradigm for aggregating information and boosting

MultiPhishGuard: An Explainable and Adaptive Multi-Agent LLM System for Phishing Email Detection

arXiv:2505.23803v2 Announce Type: replace-cross Abstract: Phishing email detection faces significant challenges due to evolving adversarial tactics and heterogeneous attack patterns. Traditional appro

Omissive Bias in Religious Representation: Benchmarking LLM Answers to Everyday Ethical Decision-making

arXiv:2605.24319v1 Announce Type: new Abstract: As large language models become a default source of guidance on personal, moral, and existential questions, it matters whether they draw on the religiou

PASA: A Principled Embedding-Space Watermarking Approach for LLM-Generated Text under Semantic-Invariant Attacks

arXiv:2605.10977v2 Announce Type: replace-cross Abstract: Watermarking for large language models (LLMs) is a promising approach for detecting LLM-generated text and enabling responsible deployment. Ho

Applications

SeqRoute: Global Budget-Aware Sequential LLM Routing via Offline Reinforcement Learning

arXiv:2605.25424v1 Announce Type: cross Abstract: Existing LLM routing frameworks treat queries as independent events, neglecting the sequential nature of real-world user sessions constrained by globa

applicationsarxiv-cs-ai

TIAR: Trajectory-Informed Advantage Reweighting for LLM Abstention Learning

arXiv:2605.25850v1 Announce Type: cross Abstract: This paper investigates large language model (LLM) abstention learning, specifically using ternary reward, which incentivize truthfulness in large lan

Truthful Online Preference Aggregation for LLM Fine-Tuning in Mobile Crowdsourcing

arXiv:2605.24052v1 Announce Type: cross Abstract: To better serve users' demands in mobile applications (e.g., navigation), mobile crowdsourcing platforms can iteratively align large language model (L

Uncovering Vulnerabilities of LLM-Assisted Cyber Threat Intelligence

arXiv:2509.23573v4 Announce Type: replace-cross Abstract: Large language models (LLMs) are increasingly used to help security analysts manage the surge of cyber threats, automating tasks from vulnerab

CapTrack: Multifaceted Evaluation of Forgetting in LLM Post-Training

arXiv:2603.06610v2 Announce Type: replace Abstract: Large language model (LLM) post-training enhances latent skills, unlocks value alignment, improves performance, and enables domain adaptation. Unfor

safetyarxiv-cs-lg

Instance-Optimal Estimation with Multiple LLM Judges on a Budget

arXiv:2605.23362v1 Announce Type: new Abstract: Evaluating large language models increasingly relies on LLM-as-a-judge protocols, but such evaluations remain costly: different judges have different pr

researcharxiv-cs-lg

Robust LLM Watermarking with Minimal Semantic Distortion for IP Protection

arXiv:2605.23175v1 Announce Type: cross Abstract: Proprietary large language models (LLMs) face risks of intellectual property (IP) violation, as adversaries can replicate an LLM by collecting input-o

researcharxiv-cs-cl

When Planning Fails Despite Correct Execution: On Epistemic Calibration for LLM-Based Multi-Agent Systems

arXiv:2605.23414v1 Announce Type: new Abstract: LLM-based multi-agent systems can fail even when planned actions are executed correctly because agents may misjudge their knowledge when evaluating plan

LLM companies are likely to be like airline companies: small margins, intense competition, high expenses.

LLM companies are likely to be like airline companies: small margins, intense competition, high expenses. @GaryMarcus Airlines taught us something important: unlimited demand does not guarantee attrac

safetygary-marcus--x

24 May 2026

not using LLM’s works for me

not using LLM’s works for me Prolonged AI use may make it harder to think critically and creatively, recent research suggests. But there are ways to keep the brain fit https://www.economist.com/scienc

safetygary-marcus--x

24 May 2026

GraphFlow: A Graph-Based Workflow Management for Efficient LLM-Agent Serving

arXiv:2605.22566v1 Announce Type: new Abstract: Large Language Model (LLM)-based agents demonstrate strong reasoning and execution capabilities on complex tasks when guided by structured instructions,

oh even better, since you can test the runtime without an LLM, you can post this into Claude: “look up http://activegraph.ai, install and ru…

model-releasesyohei-nakajima--x

oh even better, since you can test the runtime without an LLM, you can post this into Claude: “look up http://activegraph.ai, install and run a small experiment I would like leveraging this, and expla

Local Ai

OPPO: Bayesian Value Recursion for Token-Level Credit Assignment in LLM Reasoning

arXiv:2605.21851v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards has become the standard recipe for improving LLM reasoning, but the dominant algorithm GRPO assigns a sin

local-aiarxiv-cs-lg

Tailoring Teaching to Aptitude: Direction-Adaptive Self-Distillation for LLM Reasoning

arXiv:2605.22263v1 Announce Type: new Abstract: On-policy self-distillation (OPSD) is an emerging LLM post-training paradigm in which the model serves as its own teacher: conditioned on privileged inf

safetyarxiv-cs-lg

An Application-Layer Multi-Modal Covert-Channel Reference Monitor for LLM Agent Egress

arXiv:2605.20734v1 Announce Type: cross Abstract: A large language model (LLM) agent that sends messages can leak data inside them. Destination allowlists and content scanners do not police whether an

Blind Spots in the Guard: How Domain-Camouflaged Injection Attacks Evade Detection in Multi-Agent LLM Systems

arXiv:2605.22001v1 Announce Type: cross Abstract: Injection detectors deployed to protect LLM agents are calibrated on static, template-based payloads that announce themselves as override directives.

Causal Past Logic for Runtime Verification of Distributed LLM Agent Workflows

arXiv:2605.20923v1 Announce Type: cross Abstract: Distributed LLM agent workflows should not be monitored as if they produced a single sequential log. In an asynchronous execution, a decision can only

CTFExplorer: Evaluating LLM Offensive Agents Through Multi-Target Web CTF Benchmarking

arXiv:2602.08023v3 Announce Type: replace-cross Abstract: Existing benchmarks for LLM-based offensive security agents use isolated, single-target setups with a known vulnerable service and fixed objec

Diagnosis Is Not Prescription: Linguistic Co-Adaptation Explains Patching Hazards in LLM Pipelines

arXiv:2605.21958v1 Announce Type: new Abstract: When a multi-module LLM agent fails, the module most responsible for the failure is not necessarily the best place to intervene. We demonstrate this Dia

safetyarxiv-cs-cl

From Reasoning Chains to Verifiable Subproblems: Curriculum Reinforcement Learning Enables Credit Assignment for LLM Reasoning

arXiv:2605.22074v1 Announce Type: cross Abstract: Reinforcement learning from verifiable rewards (RLVR) has shown strong promise for LLM reasoning, but outcome-based RLVR remains inefficient on hard p

researcharxiv-cs-cl

Stdlib or Third-Party? Empirical Performance and Correctness of LLM-Assisted Zero-Dependency Python Libraries

arXiv:2605.21405v1 Announce Type: cross Abstract: Third-party Python libraries introduce dependency management overhead, supply chain risk, and deployment friction in constrained environments. A natur

APEX: Autonomous Policy Exploration for Self-Evolving LLM Agents

arXiv:2605.21240v1 Announce Type: new Abstract: LLM agents have shown strong performance across a wide range of complex tasks, including interactive environments that require long-horizon decision mak

If this is true, using the best public estimates we have of LLM resource use, solving this Erdos problem took 0.6–6.3 kWh of electricity and…

model-releasesethan-mollick--x

If this is true, using the best public estimates we have of LLM resource use, solving this Erdos problem took 0.6–6.3 kWh of electricity and about 3–31 liters of water. So that is less than three almo

Insights Generator: Systematic Corpus-Level Trace Diagnostics for LLM Agents

arXiv:2605.21347v1 Announce Type: cross Abstract: Diagnosing failures in LLM agents remains largely manual. Practitioners inspect a small subset of execution traces, form ad-hoc hypotheses, and iterat

agentsarxiv-cs-lg

Tutorials

Introspective X Training: Feedback Conditioning Improves Scaling Across all LLM Training Stages

arXiv:2605.20285v1 Announce Type: new Abstract: We tackle the question of how to scale more efficiently across the many, ever-growing stages of current LLM training pipelines. Our guiding intuition st

tutorialsarxiv-cs-lg

Tutorials

LLM training is built on fast MatMuls. But many surrounding ops still run as memory-bound kernels. CODA reparameterizes them to hide in the …

tutorialsjeremy-howard--x

LLM training is built on fast MatMuls. But many surrounding ops still run as memory-bound kernels. CODA reparameterizes them to hide in the matmul’s shadow, fused into its epilogue before results leav

MemGym: a Long-Horizon Memory Environment for LLM Agents

arXiv:2605.20833v1 Announce Type: new Abstract: Memory is a central capability for LLM agents operating across long-horizon tasks. Existing memory benchmarks predominantly evaluate retention of person