Search: “llm”

Reconstructing the Right Episode: Evaluating Interleaved Conversational Memory Beyond Long Context

arXiv:2608.25655v1 Announce Type: new Abstract: Conversations with chat assistants increasingly span many topics in a single long-running thread, challenging memory systems. Existing long-context and

27 Aug 2026

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Model Releases

Unmatched Does Not Mean False: Incomplete Reference Sets Can Reverse Calibration Rankings in Open-Ended Theory-of-Mind Tracking

DGX agent

arXiv:2608.25654v1 Announce Type: new Abstract: Open-ended Theory-of-Mind (ToM) trackers emit valid beliefs absent from finite references. A finite-reference-plus-matcher pipeline marks unmatched outp

model-releasesarxiv-cs-cl

27 Aug 2026

AHEAD: Adaptive Hindsight with Environment-Augmented Distillation for Agentic RL

arXiv:2608.24114v1 Announce Type: new Abstract: Training multi-turn LLM agents with reinforcement learning typically relies on trajectory-level rewards, which assign a uniform advantage to every step

Applications

Beyond Accuracy: A Dual-Judge Evaluation Protocol for Vision-Language Models in Legally Grounded Tasks

arXiv:2608.24258v1 Announce Type: new Abstract: AI systems are increasingly evaluated for legally accountable settings, where correct outputs must also be justifiable against an applicable legal stand

applicationsarxiv-cs-ai

Beyond Executable Models: The Pufibara Agent Harness and the Modelica Agent Workflow Benchmark for Physical System Modeling

arXiv:2608.23653v1 Announce Type: cross Abstract: AI agents are increasingly used for simulation-driven engineering. Physical system modeling presents different requirements from general-purpose code

CoMMa: Contribution-Aware Medical Multi-Agents for Decentralized Oncology Decision Support

arXiv:2602.09159v2 Announce Type: replace Abstract: Recent multi-agent frameworks have shown promise for oncology decision support, yet most assume centralized data access and rely on prompt-based ass

Constraint-Guided Enterprise Data Mapping with Large Language Models

arXiv:2608.24218v1 Announce Type: new Abstract: Enterprise entity alignment must handle semi-structured records, implicit attributes, and unit or granularity mismatches. Manual matching is still commo

Low-Rank Ternary Adaptation for Fine-Tuning Transformers

model-releasesarxiv-cs-cv

arXiv:2608.24469v1 Announce Type: new Abstract: Ternary transformers offer extreme memory and compute efficiency, but existing low-bit LoRA-based methods cannot directly fine-tune ternary weights. Cur

Neurosymbolic Alignment for Physiologically-Safe Clinical Language Models

arXiv:2608.24534v1 Announce Type: new Abstract: Clinical LLMs can generate recommendations that are factually plausible yet physiologically unsafe. We investigate whether safety alignment can be impro

ResiSpec: Enhancing Multi-Candidate Speculative Sampling via Residual Distribution Shaping

arXiv:2608.24411v1 Announce Type: new Abstract: The efficiency of Large Language Model (LLM) serving is fundamentally limited by the sequential nature of autoregressive decoding. Speculative Decoding

ToolRobustBench: Stage-Wise Perturbation Evaluation and Failure Diagnosis for Tool-Calling Agents

arXiv:2608.23635v1 Announce Type: cross Abstract: Large language models (LLMs) rely on tool calling as a fundamental agent capability, enabling them to invoke external systems and complete tasks beyon

VisCache: Visual KV Cache Pruning for Efficient Vision Large Language Model Inference

arXiv:2608.24063v1 Announce Type: cross Abstract: While Vision Large Language Models (VLLMs) have achieved remarkable success in multimodal reasoning, their long-context inference remains prohibitivel

BanglaVeilGuard: Cross-Script Safety Benchmarking and Lightweight Guardrails for Bangla Large Language Models

arXiv:2608.21880v1 Announce Type: new Abstract: Bangla large language model (LLM) safety is difficult to evaluate with English-centric or standard-script benchmarks because Bangla users routinely writ

FCPRAG: Fusion-Controller Parametric Retrieval-Augmented Generation for Stable Multi-Passage LoRA Injection

arXiv:2608.21750v1 Announce Type: new Abstract: Parametric retrieval-augmented generation (PRAG) injects retrieved evidence into a large language model (LLM) through passage-specific LoRA adapters, re

researcharxiv-cs-cl

HiDiffTIR: Hierarchical Difficulty-Aware Policy Optimization for Multi-Turn Tool-Integrated Reasoning

arXiv:2608.21863v1 Announce Type: cross Abstract: Tool-Integrated Reasoning (TIR) is a fundamental capability for LLM agents to solve complex tasks by interacting with external tools iteratively. Rein

Safety-Aligned Weights Are Not Enough: Refusal-Teacher-Guided Finetuning Enhances Safety and Downstream Performance under Harmful Finetuning Attacks

arXiv:2506.07356v3 Announce Type: replace Abstract: While Finetuning-as-a-Service (FaaS) enables customization of Large Language Models (LLMs) using user data, this service is vulnerable to safety deg

safetyarxiv-cs-cl

The Interaction Tax: When Communication Erases Diversity in Multi-Agent Teams

arXiv:2608.23541v1 Announce Type: cross Abstract: Does multi-agent LLM interaction help or hurt? Some work reports gains from debate (Du et al., 2024), critique loops (Chen et al., 2025), and mixture-

Asymmetric Capacity Allocation in Self-Refinement Pipelines

model-releasesarxiv-cs-lg

arXiv:2608.21345v1 Announce Type: new Abstract: Self-refinement, typically structured as generation, critique, and revision, is a widely adopted paradigm for improving LLM generation and serves as a c

Categorical AI phenomenology: A first-person approach

arXiv:2608.20420v1 Announce Type: new Abstract: This paper develops a phenomenology-first approach to artificial consciousness by reframing consciousness as the subjective experience enacted through a

Intent Engine: Natural-Language Intent Translation for Intent-Driven Orchestration in the Compute Continuum

arXiv:2608.20388v1 Announce Type: new Abstract: Microservice placement in the compute continuum is driven by low-level Service-level Objectives (SLOs), but requiring users to specify metric-level cons

Large Language Models at the Intersection of Software Engineering and Software Security:An Evidence-Centered Structured Survey and Research Agenda

arXiv:2608.21107v1 Announce Type: new Abstract: Large Language Models (LLMs) are moving from code completion toward repository-scale agents that retrieve context, edit files, execute tools, and partic

Structure for Reading, Prose for Writing: Asymmetric Structural Conditioning in Multi-Agent Document Authoring

arXiv:2608.20786v1 Announce Type: new Abstract: Multi-agent pipelines that author formal documents must both read a requester's forms and write against them. We report a deployed tender-response syste

Applications

StruProKGR: A Structural and Probabilistic Framework for Sparse Knowledge Graph Reasoning

arXiv:2512.12613v2 Announce Type: replace Abstract: Sparse Knowledge Graphs (KGs) are commonly encountered in real-world applications, where knowledge is often incomplete or limited. Sparse KG reasoni

applicationsarxiv-cs-cl

Hardware

Valid Inference with Synthetic Data via Task Exchangeability

arXiv:2606.13629v2 Announce Type: replace-cross Abstract: There is a proliferation of work arguing for the use of synthetic data in scientific research. For example, social scientists are arguing for

hardwarearxiv-cs-ai

When Words Are Safe But Actions Kill: Probing Physical Jailbreak Beyond Textual Jailbreak in Hidden-State Risk Space

arXiv:2607.15218v2 Announce Type: replace Abstract: Large language models (LLMs) increasingly serve as high-level planners for embodied agents, where linguistically benign instructions can become unsa

Forking Fast: Efficiently Estimating Uncertainty Dynamics in Text Generation

arXiv:2608.19611v1 Announce Type: cross Abstract: LLM reasoning is stochastic, and so understanding a model requires grappling with the distribution of reasoning chains that it might produce for a giv

Pandora's AI Model Routing Box: Efficient Allocation with Costly Value Estimation

arXiv:2608.20316v1 Announce Type: new Abstract: Heterogeneous AI systems composed of multiple models, architectures, harnesses, or inference-time settings can improve quality and efficiency by routing

Scale-Aware Pretraining of Time Series Foundation Models via Multi-Patch Token Alignment and Hybrid Masking

arXiv:2608.20005v1 Announce Type: new Abstract: Pretraining time series foundation models across heterogeneous datasets necessitates effective handling of varying sampling frequencies. Current methods

safetyarxiv-cs-lg

Towards Audio Token Compression in Large Audio Language Models

arXiv:2511.20973v2 Announce Type: replace-cross Abstract: Large Audio Language Models (LALMs) deliver strong performance across speech and audio tasks, but their audio encoders generate high-rate toke

Tutorials

What You Can't See Is What You Learn: Restricted Evidence Visibility Favors Compositional Generalization in Shared-Genome Language-Model Societies

arXiv:2608.20054v1 Announce Type: new Abstract: Multi-module systems often expose every module to the full input. We test whether restricting evidence visibility changes which solutions gradient-based

tutorialsarxiv-cs-ai

FraudBench: Stress-Testing Policy-Grounded Banking Agents Against Adaptive Fraud

arXiv:2608.18136v1 Announce Type: new Abstract: Conversational agents now act for end users through tools while holding access to customer databases and internal policy documents that a caller can rea

Hallucination Detection in Large Language Models Using Diversion Decoding

arXiv:2607.10476v2 Announce Type: replace Abstract: Large language models (LLMs) have emerged as a powerful tool for retrieving knowledge through seamless, human-like interactions. Despite their advan

researcharxiv-cs-cl

Institutional Prestige as Geographic Bias in Large Language Models: Evidence from Three Factorial Experiments with Bootstrap Confidence Intervals

arXiv:2608.18107v1 Announce Type: cross Abstract: We investigate whether large language models (LLMs) systematically discriminate in candidate evaluations based on applicant name ethnicity and/or inst

SkillGate: Training In-Policy Skill Selection in Long-Horizon Agents

arXiv:2608.18852v1 Announce Type: new Abstract: Agent frameworks increasingly package procedural knowledge as skills: instruction files an agent reads on demand, while public libraries now hold thousa

StocksTalk: A Voice-Enabled Conversational Agent for Structured Query Generation over Web Data

arXiv:2608.18105v1 Announce Type: cross Abstract: StocksTalk is a voice-enabled conversational system for transforming spoken financial screening requests into executable and validated structured quer

Agent Lightning v1.0: Towards Harnessed Agentic RL

arXiv:2608.17528v1 Announce Type: new Abstract: Modern agents operate inside agent harnesses that manage tools, context, and control flow, making the harness a critical part of the agent system. Our o

CABLE: Extending the Reach of Memory Retrieval via Complementary Antecedent-Based Linking and Expansion

arXiv:2608.17911v1 Announce Type: new Abstract: As LLM agents operate across structured workflows and sessions, preserving long-term history does not ensure that later contexts can recover relevant ev

DataSTORM: Deep Research on Large-Scale Databases using Exploratory Data Analysis and Data Storytelling

arXiv:2604.06474v2 Announce Type: replace Abstract: Deep research with Large Language Model (LLM) agents is emerging as a powerful paradigm for multi-step information discovery, synthesis, and analysi

agentsarxiv-cs-cl

Effects of Answer Format Variation on Gender Bias in Large Language Models

arXiv:2608.17516v1 Announce Type: new Abstract: Gender bias or other social biases in large language models (LLMs) are frequently evaluated with question answering or survey benchmarks where the LLM n

Fair ASR: Re-Evaluating Black-Box Jailbreaks under Shared Target-Call Budgets

arXiv:2608.17360v1 Announce Type: cross Abstract: Reliable jailbreak evaluation is essential for assessing LLM safety, but most existing studies rely solely on attack success rate (ASR) without accoun

MobileWorldSafety: Benchmarking GUI Agent Safety Against Environmental Injection Attacks in Android Apps

arXiv:2608.17659v1 Announce Type: cross Abstract: LLM-powered GUI agents that autonomously operate smartphones are rapidly transitioning from research prototypes to early real-world deployment. Howeve

On the Fragility of Self-Improving Agents: Variance, Task Order, and Underspecification

arXiv:2608.18066v1 Announce Type: new Abstract: Memory-based self-improving agents--those that learn from an online stream of tasks and improve over time by maintaining a textual memory bank--have sho

PACE: Policy-Attested Contract Execution for Safe AI Agents in Decentralized Finance

arXiv:2608.17220v1 Announce Type: cross Abstract: Autonomous AI agents are emerging as interfaces for decentralized finance (DeFi) actions such as swaps, lending operations, and yield management. Beca

PrivAct: Internalizing Contextual Privacy Preservation via Multi-Agent Preference Training

arXiv:2602.13840v2 Announce Type: replace Abstract: Large language model (LLM) agents are increasingly deployed in personalized tasks involving sensitive, context-dependent information, where privacy

agentsarxiv-cs-cl

Runtime Governance for Agentic AI: Action-Boundary Control with Trusted Provenance and Fail-Closed Execution

arXiv:2608.16891v1 Announce Type: new Abstract: Agentic AI systems request tool actions that can modify files, send messages, launch jobs, or change workflow state. This shifts the safety problem from

TRUSS: Towards Task-Reliable and User-Safe Automated Agent Skill Generation

arXiv:2608.17588v1 Announce Type: new Abstract: Agent Skills package reusable natural language procedures with executable resources, enabling software agents to acquire task specific capabilities with

When Agents Act on Web3: An Attack-Surface Survey of MCP, Skills, and Tool Calling

arXiv:2608.17275v1 Announce Type: cross Abstract: AI agents increasingly act rather than merely read: across the Model Context Protocol (MCP) ecosystem, the share of deployed tools that modify externa

AutoMem: A Text-Gradient Recursive Self-Improvement Framework for Automated Memory Architectures Search

arXiv:2608.14621v1 Announce Type: cross Abstract: Long-term memory is increasingly central to LLM agents, yet memory design remains a highly coupled architecture problem: what to encode, how to store