The Morning
From the arXiv
AgentHPOBench: A Benchmark For Evaluating LLM Agents as Sequential Hyperparameter Optimizers
his paper introduces AgentHPOBench, a novel benchmark designed to evaluate LLM agents' ability to perform sequential hyperparameter optimization. Unlike previous benchmarks, it assesses how agents interpret experimental evidence to guide subsequent configuration choices across diverse machine learning tasks. The contribution lies in providing a standardized framework to measure and compare the experimental optimization capabilities of LLM agents.
AMTFV: Agentic Mathematical Tool-Flow Verification for LLM Self-Correction
AMTFV introduces a novel "Mathematical Tool Flow" (MTF) interface to enable LLMs to reliably verify their mathematical answers. This method decouples verification modeling from execution by allowing the LLM to construct a workflow, request specific computation…
Beyond Component Testing: Validating Agentic AI Systems
This paper addresses the challenge of validating complex agentic AI systems, which exhibit multi-step, dynamic behaviors. It synthesizes existing research to propose a five-dimension taxonomy (behavioral, safety, temporal, regulatory, multi-agent) for characte…
LEMUR: Learning to Align with Multi-Objective Reinforcement Learning from Preference Feedback
LEMUR addresses the challenge of training RL agents for tasks with multiple, conflicting objectives when explicit reward functions are unavailable. It combines Multi-Objective Reinforcement Learning with Preference-based RL, enabling agents to learn complex tr…
Tool Specifications Matter: Uncovering and Mitigating Safety Risks in AI Agents
This paper argues that the way AI agents are given instructions for using external tools (tool specifications) significantly impacts their safety. They found that schema-formatted specifications weaken the AI's ability to refuse harmful actions. To address thi…
Knowing When to Quit: Diagnosing and Training LLMs to Abort Futile Reasoning
This paper introduces **CaRL**, a method to train Large Language Models (LLMs) to recognize and abort "futile reasoning" on tasks exceeding their capabilities. CaRL uses reward sha…
TokTier: Exact Stateful Tokenization for Agentic LLM Serving
TokTier addresses the inefficiency of re-tokenizing entire prompts in agentic LLM serving. Its core method is stateful tokenization that guarantees identical token IDs to full refe…
Zero-Mem: Zero-Token Memory Operations for LLM Agents
Zero-Mem proposes a novel approach to LLM agent memory by eliminating token costs for memory operations. Instead of using LLM calls, it organizes interaction traces into an entity-…
AgenticRepair: Multi-Faceted Program Context Engineering for Agentic Vulnerability Repair
This paper introduces AgenticRepair, a framework for automated vulnerability repair. Its core method is multi-faceted program context engineering, which addresses critical gaps in …
Beyond Retrieval: Analytic Memory for Multimodal Agents
This paper introduces "analytic memory" as a new paradigm for multimodal agents, complementing existing "retrieval memory." Analytic memory allows agents to compute over accumulate…
The Town Square
Qwen3.8-Max, a new large language model, sets a higher standard for coding assistance and collaborative tasks.
Workshops
This open-source project management tool, Kaneo, aims to be a comprehensive yet streamlined solution that adapts to user needs, focusing on efficiency and usability.
AirLLM enables efficient inference of the 70B LLM on a single 4GB GPU by optimizing memory usage and computational resources.
