Search: “llm”

SpecBlock: Block-Iterative Speculative Decoding with Dynamic Tree Drafting

arXiv:2605.07243v1 Announce Type: new Abstract: Speculative decoding accelerates LLM inference by drafting a tree of candidate continuations and verifying it in one target forward. Existing drafters f

Hardware

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

SpikingBrain: Spiking Brain-inspired Large Models

DGX agent

arXiv:2509.05276v4 Announce Type: replace-cross Abstract: Mainstream Transformer-based large language models face major efficiency bottlenecks: training computation scales quadratically with sequence

hardwarearxiv-cs-ai

11 May 2026

TRACE: Tourism Recommendation with Accountable Citation Evidence

arXiv:2605.07677v1 Announce Type: cross Abstract: Tourism is a high-stakes setting for conversational recommender systems (CRS): a plausible-sounding suggestion can waste real money and trip time once

TraceFix: Repairing Agent Coordination Protocols with TLA+ Counterexamples

arXiv:2605.07935v1 Announce Type: new Abstract: We present TraceFix, a verification-first pipeline for Large Language Model (LLM) multi-agent coordination. An agent synthesizes a protocol topology as

Understanding Performance Collapse in Layer-Pruned Large Language Models via Decision Representation Transitions

arXiv:2605.07271v1 Announce Type: cross Abstract: Layer pruning efficiently reduces Large Language Model (LLM) computational costs but often triggers sudden performance collapse. Existing representati

When Routine Chats Turn Toxic: Unintended Long-Term State Poisoning in Personalized Agents

arXiv:2605.06731v1 Announce Type: cross Abstract: Personalized LLM agents maintain persistent cross-session state to support long-horizon collaboration. Yet, this persistence introduces a subtle but c

Agentic publications: redesigning scientific publishing in the age of thinking large language models

arXiv:2505.13246v2 Announce Type: replace Abstract: Purpose: This paper introduces the concept of 'Agentic Publication,' a novel LLM-driven framework designed to complement traditional scientific publ

Conceptors for Semantic Steering

arXiv:2605.04980v1 Announce Type: cross Abstract: Activation-based steering provides control of LLM behavior at inference time, but the dominant paradigm reduces each concept to a single direction who

Consensus Entropy: Harnessing Multi-VLM Agreement for Self-Verifying and Self-Improving OCR

arXiv:2504.11101v4 Announce Type: replace Abstract: Optical Character Recognition (OCR) is fundamental to Vision-Language Models (VLMs) and high-quality data generation for LLM training. Yet, despite

researcharxiv-cs-cv

EP-GRPO: Entropy-Progress Aligned Group Relative Policy Optimization with Implicit Process Guidance

arXiv:2605.04960v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards (RLVR), particularly Group Relative Policy Optimization (GRPO), has advanced LLM reasoning. However, GRPO

safetyarxiv-cs-lg

Self-Induced Outcome Potential: Turn-Level Credit Assignment for Agents without Verifiers

arXiv:2605.04984v1 Announce Type: cross Abstract: Long-horizon LLM agents depend on intermediate information-gathering turns, yet training feedback is usually observed only at the final answer, becaus

agentsarxiv-cs-cl

What Happens Inside Agent Memory? Circuit Analysis from Emergence to Diagnosis

arXiv:2605.03354v1 Announce Type: new Abstract: Agent memory failures are silent: an LLM-based agent can produce a fluent response even when it fails to extract, retain, or retrieve the information ne

APIOT: Autonomous Vulnerability Management Across Bare-Metal Industrial OT Networks

arXiv:2605.02346v1 Announce Type: cross Abstract: Bare-metal operational technology (OT) devices -- especially the microcontrollers running Modbus/TCP and CoAP at the base of industrial control system

ExaGPT: Example-Based Machine-Generated Text Detection for Human Interpretability

arXiv:2502.11336v2 Announce Type: replace Abstract: Detecting texts generated by Large Language Models (LLMs) could cause grave mistakes due to incorrect decisions, such as undermining students' acade

Hybrid Inspection and Task-Based Access Control in Zero-Trust Agentic AI

arXiv:2605.02682v1 Announce Type: new Abstract: Authorizing Large Language Model (LLM)-driven agents to dynamically invoke tools and access protected resources introduces significant security risks, a

MEMSAD: Gradient-Coupled Anomaly Detection for Memory Poisoning in Retrieval-Augmented Agents

arXiv:2605.03482v1 Announce Type: cross Abstract: Persistent external memory enables LLM agents to maintain context across sessions, yet its security properties remain formally uncharacterized. We for

OpenSeeker-v2: Pushing the Limits of Search Agents with Informative and High-Difficulty Trajectories

arXiv:2605.04036v1 Announce Type: cross Abstract: Deep search capabilities have become an indispensable competency for frontier Large Language Model (LLM) agents, yet their development remains dominat

agentsarxiv-cs-cl

Position: How can Graphs Help Large Language Models?

arXiv:2605.02452v1 Announce Type: new Abstract: With the rapid advancement of large language models (LLMs), classic graph learning tasks have greatly benefited from LLMs, including improved encoding o

Prism: Efficient Test-Time Scaling via Hierarchical Search and Self-Verification for Discrete Diffusion Language Models

model-releasesarxiv-cs-lg

arXiv:2602.01842v3 Announce Type: replace Abstract: Inference-time compute has re-emerged as a practical way to improve LLM reasoning. Most test-time scaling (TTS) algorithms rely on autoregressive de

Separating Intelligence from Execution: A Workflow Engine for the Model Context Protocol

arXiv:2605.00827v1 Announce Type: cross Abstract: Large Language Model (LLM) agents increasingly interact with external systems through tool-calling protocols such as the Model Context Protocol (MCP).

To Write or to Automate Linguistic Prompts, That Is the Question

arXiv:2603.25169v2 Announce Type: replace Abstract: LLM performance is highly sensitive to prompt design, yet whether automatic prompt optimization can replace expert prompt engineering in linguistic

Towards Multi-Agent Autonomous Reasoning in Hydrodynamics

arXiv:2605.01102v1 Announce Type: new Abstract: Single-agent systems (SAS) have become the default pattern for LLM-driven scientific workflows, but routing planning, tool use, and synthesis through a

Towards Understanding Specification Gaming in Reasoning Models

arXiv:2605.02269v1 Announce Type: new Abstract: Specification gaming is a critical failure mode of LLM agents. Despite this, there has been little systematic research into when it arises and what driv

Trojan Hippo: Weaponizing Agent Memory for Data Exfiltration

arXiv:2605.01970v2 Announce Type: cross Abstract: Memory systems enable otherwise-stateless LLM agents to persist user information across sessions, but also introduce a new attack surface. We characte

Hardware

ZeRO-Prefill: Zero Redundancy Overheads in MoE Prefill Serving

arXiv:2605.02960v1 Announce Type: new Abstract: Production LLM workloads increasingly serve discriminative tasks, such as classification, recommendation, and verification, whose answers are read from

hardwarearxiv-cs-lg

AEM: Adaptive Entropy Modulation for Multi-Turn Agentic Reinforcement Learning

arXiv:2605.00425v1 Announce Type: new Abstract: Reinforcement learning (RL) has significantly advanced the ability of large language model (LLM) agents to interact with environments and solve multi-tu

Benchmarking Retrieval Strategies for Biomedical Retrieval-Augmented Generation: A Controlled Empirical Study

arXiv:2605.02520v1 Announce Type: new Abstract: Retrieval-Augmented Generation (RAG) offers a well-established path to grounding large language model (LLM) outputs in external knowledge, yet the quest

DynamicPO: Dynamic Preference Optimization for Recommendation

arXiv:2605.00327v1 Announce Type: cross Abstract: In large language model (LLM)-based recommendation systems, direct preference optimization (DPO) effectively aligns recommendations with user preferen

EditPropBench: Measuring Factual Edit Propagation in Scientific Manuscripts

arXiv:2605.02083v1 Announce Type: new Abstract: Local factual edits in scientific manuscripts often create non-local revision obligations. If a dataset changes from 215 to 80 documents, claims such as

Feedback-Normalized Developer Memory for Reinforcement-Learning Coding Agents: A Safety-Gated MCP Architecture

arXiv:2605.01567v1 Announce Type: cross Abstract: Large language model (LLM) coding agents increasingly operate over repositories, terminals, tests, and execution traces across long software-engineeri

InfoLaw: Information Scaling Laws for Large Language Models with Quality-Weighted Mixture Data and Repetition

arXiv:2605.02364v1 Announce Type: new Abstract: Upweighting high-quality data in LLM pretraining often improves performance, but in datalimited regimes, especially under overtraining, stronger upweigh

Hardware

Maistros: A Greek Large Language Model Adapted Through Knowledge Distillation From Large Reasoning Models

arXiv:2605.01870v1 Announce Type: new Abstract: Large Language Models (LLMs) have substantially advanced the field of Natural Language Processing (NLP), achieving state-of-the-art performance across a

hardwarearxiv-cs-cl

Minimizing Collateral Damage in Activation Steering

arXiv:2605.01167v1 Announce Type: new Abstract: Activation steering is a method for controlling Large Language Model (LLM) behavior by intervening in its internal representations to increase the align

safetyarxiv-cs-lg

Molecular Representations for Large Language Models

model-releasesarxiv-cs-lg

arXiv:2605.01822v1 Announce Type: new Abstract: Large Language Models (LLMs) are increasingly being used to support scientific discovery. In chemistry, tasks such as reaction prediction and structure

Q-RAG: Long Context Multi-step Retrieval via Value-based Embedder Training

arXiv:2511.07328v2 Announce Type: replace Abstract: Retrieval-Augmented Generation (RAG) methods enhance LLM performance by efficiently filtering relevant context for LLMs, reducing hallucinations and

Skills as Verifiable Artifacts: A Trust Schema and a Biconditional Correctness Criterion for Human-in-the-Loop Agent Runtimes

arXiv:2605.00424v1 Announce Type: cross Abstract: Agent skills -- structured packages of instructions, scripts, and references that augment a large language model (LLM) without modifying the model its

safetyarxiv-cs-ai

SpecKV: Adaptive Speculative Decoding with Compression-Aware Gamma Selection

arXiv:2605.02888v1 Announce Type: cross Abstract: Speculative decoding accelerates large language model (LLM) inference by using a small draft model to propose candidate tokens that a larger target mo

Ultrasound Vision-Language Alignment via Contrastive Learning

arXiv:2605.02126v1 Announce Type: new Abstract: Ultrasound foundation models have achieved strong performance on structured prediction tasks but remain exclusively vision-based, limiting zero-shot and

safetyarxiv-cs-cv

Applications

VAUQ: Vision-Aware Uncertainty Quantification for LVLM Self-Evaluation

arXiv:2602.21054v2 Announce Type: replace-cross Abstract: Large Vision-Language Models (LVLMs) frequently hallucinate, limiting their safe deployment in real-world applications. Existing LLM self-eval

applicationsarxiv-cs-cl

When RL Meets Adaptive Speculative Training: A Unified Training-Serving System

model-releasesarxiv-cs-lg

arXiv:2602.06932v3 Announce Type: replace Abstract: Speculative decoding can significantly accelerate LLM serving, yet most deployments today disentangle speculator training from serving, treating spe

Certain Head, Uncertain Tail: Expert-Sample for Test-Time Scaling in Fine-Grained MoE

arXiv:2602.02443v2 Announce Type: replace Abstract: Test-time scaling improves LLM performance by generating multiple candidate solutions, yet token-level sampling requires temperature tuning that tra

Decouple before Integration: Test-time Synthesis of SFT and RLVR Task Vectors

arXiv:2605.00610v1 Announce Type: new Abstract: SFT and RLVR represent two fundamental yet distinct paradigms for LLM post-training, each excelling in distinct dimensions. SFT expands knowledge breadt

Applications

Evaluating Legal Reasoning Traces with Legal Issue Tree Rubrics

arXiv:2512.01020v2 Announce Type: replace-cross Abstract: Evaluating the quality of LLM-generated reasoning traces in expert domains (e.g., law) is essential for ensuring credibility and explainabilit

applicationsarxiv-cs-cl

Image Score: Learning and Evaluating Human Preferences for Mercari Search

arXiv:2408.11349v2 Announce Type: replace Abstract: Mercari is the largest C2C e-commerce marketplace in Japan, having more than 20 million active monthly users. Search being the fundamental way to di

researcharxiv-cs-cv

Learning How and What to Memorize: Cognition-Inspired Two-Stage Optimization for Evolving Memory

arXiv:2605.00702v1 Announce Type: new Abstract: Large language model (LLM) agents require long-term user memory for consistent personalization, but limited context windows hinder tracking evolving pre

safetyarxiv-cs-cl

PORTool: Importance-Aware Policy Optimization with Rewarded Tree for Multi-Tool-Integrated Reasoning

arXiv:2510.26020v2 Announce Type: replace Abstract: Multi-tool-integrated reasoning enables LLM-empowered tool-use agents to solve complex tasks by interleaving natural-language reasoning with calls t

safetyarxiv-cs-cl

Representation in large language models

arXiv:2501.00885v2 Announce Type: replace Abstract: The extraordinary success of recent Large Language Models (LLMs) on a diverse array of tasks has led to an explosion of scientific and philosophical

SAGA: Workflow-Atomic Scheduling for AI Agent Inference on GPU Clusters

arXiv:2605.00528v1 Announce Type: cross Abstract: AI agents execute tens to hundreds of chained LLM calls per task, yet GPU schedulers treat each call as independent, discarding gigabytes of intermedi

safetyarxiv-cs-lg