Search: “llm”

Win by Silence: Deletion Non-Monotonicity, Autonomous Exploitation, and Typed-State Gating in LLM Plan Evaluation

arXiv:2607.12986v1 Announce Type: new Abstract: Plan evaluators can reward a strategic plan for becoming less explicit. This paper studies that failure in a staged expected-value scorer for LLM-genera

agentsarxiv-cs-ai

15 Jul 2026

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Agents

'The paper's insight connects to a broader pattern: AI agents are essentially distributed systems with unreliable components (the LLM), and …

DGX agent

'The paper's insight connects to a broader pattern: AI agents are essentially distributed systems with unreliable components (the LLM), and we should apply distributed systems patterns to them.' https

agentsyohei-nakajima--x

14 Jul 2026

From Prompts to Contracts: Harness Engineering for Auditable Enterprise LLM Agents

arXiv:2607.08028v1 Announce Type: new Abstract: Enterprise large language model (LLM) applications often begin as prototypes whose behavior is carried by prompts and retrieval context. Productization

ICDAR 2026 HIPE-OCRepair Competition on LLM-Assisted OCR Post-Correction for Historical Documents

arXiv:2607.08143v1 Announce Type: cross Abstract: We present the results of HIPE-OCRepair-2026, an ICDAR competition on LLM-assisted OCR post-correction of historical documents. OCR post-correction re

Tool-Making and Self-Evolving LLM Agents in Low-Latency Systems

arXiv:2607.08010v1 Announce Type: new Abstract: Production LLM agents often waste latency and reliability by regenerating code for the same procedural steps on every request. We replace this inference

agentsarxiv-cs-cl

Local Ai

TRACE: A Two-Channel Robust Attribution Watermark via Complementary Embeddings for LLM-Agent Trajectories

arXiv:2607.08400v1 Announce Type: cross Abstract: LLM agents reach users through resellers, who may rebrand a developer's agent or substitute a cheaper model. When provenance is disputed, attribution

local-aiarxiv-cs-ai

When the Judge Changes, So Does the Measurement: Auditing LLM-as-Judge Reliability

arXiv:2607.08535v1 Announce Type: cross Abstract: An LLM-as-judge score can move even when the candidate responses stay fixed, simply because the evaluator has changed. We treat this evaluator-replace

Workflow as Knowledge: Semantic Persistence for LLM-Mediated Workflows

arXiv:2607.08740v1 Announce Type: new Abstract: Large language model (LLM) applications increasingly use explicit workflows for tool use, retrieval, branching, checkpointing, and human approval. Exist

Operational Reframing and Approval-Framed Delegation in Multi-Agent LLM Safety

arXiv:2607.07097v1 Announce Type: new Abstract: Safety evaluations of multi-agent LLM systems often compare a direct prompt with a planner-executor pipeline and report the difference as a single 'pipe

9 Jul 2026

Great writeup from the University of Oxford. It's a taxonomy of LLM-based agent limitations. Good read for anyone shipping with agents. Benc…

Great writeup from the University of Oxford. It's a taxonomy of LLM-based agent limitations. Good read for anyone shipping with agents. Benchmark scores keep climbing, yet the same agent failures resu

model-releasesdair-ai--x

Inject or Navigate? Token-Efficient Retrieval for LLM Analysis of Transactional Legal Documents

arXiv:2607.05764v1 Announce Type: new Abstract: Answering questions over a set of transactional legal documents is most simply done by injecting the whole corpus into the LLM's context window on every

PatchOptic for Shared-State LLM Workflows with Projected Views and Verified Structured Updates

arXiv:2607.05483v1 Announce Type: cross Abstract: Agentic workflows often operate over shared, structured state. Because LLM context windows are limited, each model invocation is typically shown only

The most important part of an AI agent isn't the LLM. It's what surrounds it. 👇 The first image made it click for me. While exploring LangC…

The most important part of an AI agent isn't the LLM. It's what surrounds it. 👇 The first image made it click for me. While exploring LangChain Deep Agents, I ran the same prompt. Same model. Almost t

agentsharrison-chase--x

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

arXiv:2606.20023v2 Announce Type: replace-cross Abstract: As LLM agents increasingly select tools autonomously, their choices among tools with different privileges become safety-relevant. However, pri

WordVoice: Explicit and Decoupled Multi-Dimensional Word-Level Control for LLM-Based TTS

arXiv:2607.06461v1 Announce Type: cross Abstract: While recent Large Language Model (LLM)-based Text-to-Speech (TTS) systems have achieved remarkable naturalness, they predominantly rely on implicit e

safetyarxiv-cs-cl

CausalGame: Benchmarking Causal Thinking of LLM Agents in Games

arXiv:2607.04293v1 Announce Type: cross Abstract: Building AI Scientist agents with Large Language Models (LLMs) has recently attracted growing attention. Since scientific discovery fundamentally reli

Dictionaries, Not Darwin: Set-Level Selection Beats LLM Evolution in Scientific Equation Discovery

model-releasesarxiv-cs-lg

arXiv:2607.04108v1 Announce Type: new Abstract: Large language models are increasingly used as evolutionary engines for scientific discovery: generate candidates, select winners, feed them back as par

Tutorials

Evals in CI: How to write your LLM evals as tests with Arize Phoenix

If you're struggling to get started with evals, you're not alone. This post explains how to write LLM evals as ordinary tests in CI with Phoenix, pytest, and Vitest/Jest. The post Evals in CI: How to

tutorialsarize-ai

LLM-based Human Simulations Have Not Yet Been Reliable

arXiv:2501.08579v3 Announce Type: replace Abstract: Large Language Models (LLMs) are increasingly employed for simulating human behaviors across diverse domains. However, our position is that current

safetyarxiv-cs-cl

No Time Like the Present: Agentic Test-Time Training for LLM Agents

arXiv:2607.03441v1 Announce Type: cross Abstract: LLM agents often degrade over long episodes: as trajectories grow, they revisit explored states, repeat failed actions, and lose strategies that previ

Progressive Disclosure for LLM-Maintained Wiki Knowledge Bases: a Preregistered Ablation

arXiv:2607.04576v1 Announce Type: new Abstract: LLM agents increasingly answer questions against knowledge bases they help maintain. A common intuition holds that progressive disclosure, a compact cat

agentsarxiv-cs-cl

Hardware

This is what latency optimization looks like below the API 👇 Together ATLAS, NVIDIA Blackwell, CUDA, TensorRT-LLM, Dynamo, and custom kerne…

This is what latency optimization looks like below the API 👇 Together ATLAS, NVIDIA Blackwell, CUDA, TensorRT-LLM, Dynamo, and custom kernels all working together to make inference faster for users. P

hardwaretogether-ai--x

Your Agent's Memories Are Not Its Own: Forged Reasoning Attacks on LLM Agent Memory and Defenses

arXiv:2607.05029v1 Announce Type: cross Abstract: Persistent memory has enabled large language model (LLM) agents to store factual knowledge, prior decisions, reasoning histories, tool usage informati

agentsarxiv-cs-ai

Tutorials

ADVENT: LLM-Driven Automatic Predicate Invention for ILP

arXiv:2607.01585v1 Announce Type: cross Abstract: Predicate invention (PI), the creation of new predicates to extend the hypothesis space, remains a critical bottleneck in Inductive Logic Programming

tutorialsarxiv-cs-ai

AgenticSTS: A Bounded-Memory Testbed for Long-Horizon LLM Agents

arXiv:2607.02255v1 Announce Type: new Abstract: Memory for a long-horizon LLM agent is a contract about what each future decision is allowed to see. The simplest contract appends past observations, to

Hardware

DeadPool: Resilient LLM Training with Hot-Swapping via Zero-Overhead Checkpoint

arXiv:2607.01646v1 Announce Type: new Abstract: State-of-the-art large language model (LLM) training takes tens of thousands of graphics processing units (GPUs) for months and encounters failures acro

hardwarearxiv-cs-lg

LLM Priors for ERM over Programs

arXiv:2510.14331v3 Announce Type: replace Abstract: We study program-learning methods that are efficient in both samples and computation. Classical learning theory suggests that when the target admits

researcharxiv-cs-lg

Prompt Framing Distorts Count-Based Evaluation of LLM Error Detection: Evidence from Numeric Anchoring

arXiv:2607.01240v1 Announce Type: cross Abstract: Count-based F1 is widely used as a proxy for LLM error-detection quality, but this paper shows that it can rise dramatically without a corresponding i

Rethinking Complexity Metrics for LLM-Integrated Applications: Beyond Source Code

arXiv:2607.01903v1 Announce Type: new Abstract: LLM-integrated applications blend natural language prompts with program code, and much of their runtime behavior originates in the prompt layer rather t

researcharxiv-cs-ai

The Dual Nature of LLM Persona: Aggregated Tendencies and Frame-Dependent Geometry

arXiv:2607.02368v1 Announce Type: cross Abstract: Evaluations of LLM personas via psychometric questionnaires typically rely on aggregate scores, discarding within-instance correlation structure. We t

researcharxiv-cs-ai

Persona Non Grata: LLM Persona-Driven Generations in MCQA are Unstable in Distinct Dimensions

arXiv:2607.00937v1 Announce Type: new Abstract: Persona-driven generations (PDGs) have seen prolific use in research and industry applications, where a large language model (LLM) takes on a 'persona'

researcharxiv-cs-cl

2 Jul 2026

Calibrating the Evaluator: Does Probability Calibration Mitigate Preference Coupling in LLM Agent Feedback Loops?

arXiv:2606.31371v1 Announce Type: cross Abstract: When large language model (LLM) agents adapt their behavior through evaluator feedback, systematic evaluator biases propagate into the agent's learned

Citation Discipline in Spec-Driven Development: A Cross-Model Empirical Study of Output Determinism and Automated Hallucination Detection in LLM-Generated Code

arXiv:2606.30689v1 Announce Type: cross Abstract: Spec-Driven Development (SDD) frameworks guide Large Language Model (LLM)-powered code generation through formal specifications, yet they differ funda

FairJudge: An Adaptive, Debiased, and Consistent LLM-as-a-Judge

arXiv:2602.06625v2 Announce Type: replace Abstract: Existing LLM-as-a-Judge systems suffer from three fundamental limitations: limited adaptivity to task- and domain-specific evaluation criteria, syst

Generative Skill Composition for LLM Agents

arXiv:2606.32025v1 Announce Type: new Abstract: Recent LLM agents benefit from skills for solving complex tasks. Skills encapsulate modular packages of procedural knowledge and instructions for perfor

LLM-as-a-judge validity in physics assessment depends more on the task than the model

arXiv:2603.14732v2 Announce Type: replace-cross Abstract: As large language models (LLMs) are increasingly considered for automated assessment and feedback, understanding when LLM marking is valid is

MultiUAV-Plat: An LLM-Oriented Platform, Benchmark and Framework for Multi-UAV Collaborative Task Planning

arXiv:2606.31073v1 Announce Type: new Abstract: Large language models (LLMs) provide a promising interface for high-level robotic task planning, but their use in multi-UAV collaboration remains diffic

CLQT: A Closed-Loop, Cost-Aware, Strategy-Consistent Benchmark for Diagnostic Evaluation of LLM Portfolio-Management Agents

arXiv:2606.29771v1 Announce Type: new Abstract: LLM agents are increasingly cast as autonomous portfolio managers, and benchmarks have moved from financial question-answering to sequential trading. Ye

Collective cooperation without individual fidelity in LLM agents

arXiv:2606.30454v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly used as agents in simulations of social systems, yet it remains unclear when their behavior can be inter

Depth Exploration for LLM Decoding

arXiv:2606.29223v1 Announce Type: new Abstract: Autoregressive LLM decoding evaluates every generated token through the full layer stack, even though many tokens become predictable at intermediate dep

researcharxiv-cs-lg

EvalSafetyGap: A Hybrid Survey and Conceptual Framework for LLM Evaluation-Safety Failures

arXiv:2606.30219v1 Announce Type: new Abstract: LLM evaluation and AI safety face a shared measurement problem: benchmark scores, reward-model signals, and reported safety metrics can improve while th

Evaluating Memory in LLM Agents via Incremental Multi-Turn Interactions

arXiv:2507.05257v4 Announce Type: replace-cross Abstract: Recent benchmarks for Large Language Model (LLM) agents primarily focus on evaluating reasoning, planning, and execution capabilities, while a

Governance Decay: How Context Compaction Silently Erases Safety Constraints in Long-Horizon LLM Agents

arXiv:2606.22528v2 Announce Type: replace Abstract: Modern LLM agents increasingly rely on context compaction, summarization, or eviction to keep long-running sessions within a token budget. We show t

meta-pipe: An LLM-agent pipeline for end-to-end automated systematic review and meta-analysis

arXiv:2606.28363v1 Announce Type: cross Abstract: Objective: To describe the architecture and design rationale of meta-pipe, an open-source large language model (LLM)-agent pipeline that integrates th

Tutorials

Mural: Transferring LLM knowledge to image generation via Mixture-of-Transformers

arXiv:2606.29013v1 Announce Type: new Abstract: Leveraging capabilities of large language models (LLMs) in text-to-image (T2I) synthesis is an important research direction. In this work we investigate

tutorialsarxiv-cs-cv

Selective Memory Retention for Long-Horizon LLM Agents

arXiv:2606.29178v1 Announce Type: new Abstract: When does retention matter for memory-augmented LLM agents? We study this with TraceRetain, a lightweight framework for bounded external memory in froze

Local Ai

SurrogateShield: Beyond Redaction for High-Utility, Privacy-Preserving LLM Interactions

arXiv:2606.29567v1 Announce Type: cross Abstract: LLM-based assistants transmit user queries verbatim to third-party API endpoints that lie outside the user's audit or control. When those queries cont

local-aiarxiv-cs-cl

Understanding LLM Intervention Explanations in Multi-Party Human-Robot Interaction

arXiv:2606.29460v1 Announce Type: new Abstract: Large Language Models (LLMs) are increasingly embedded in social robots to support natural group interactions, yet their role in complex multi-party set

researcharxiv-cs-ro