Tag: Agent

All the articles with the tag "Agent".

• LiteWebAgent: The Open-Source Suite for VLM-Based Web-Agent Applications

  Published: 14 May, 2025 |  at  11:12 AM

  

  90.54 🤔

  LiteWebAgent is an open-source suite for VLM-based web agents that bridges the gap in production-ready solutions by offering an extensible framework with decoupled action generation and grounding, advanced planning, memory, tree search, and practical deployments via Vercel and Chrome extension.

• Agent RL Scaling Law: Agent RL with Spontaneous Code Execution for Mathematical Problem Solving

  Published: 22 May, 2025 |  at  11:12 AM

  

  89.47 🤔

  本文通过ZeroTIR框架利用强化学习训练基础大型语言模型自发执行Python代码解决数学问题，揭示了训练步数与代码使用频率、响应长度及任务准确率的正相关规律（Agent RL Scaling Law），并在数学基准上显著优于无工具基线。

• MELON: Provable Indirect Prompt Injection Defense via Masked Re-execution and Tool Comparison

  Published: 8 May, 2025 |  at  10:22 AM

  

  89.40 🤔

  MELON introduces a novel training-free defense against indirect prompt injection attacks on LLM agents by detecting independence of tool calls from user inputs through masked re-execution, achieving superior attack prevention (0.24% ASR on GPT-4o) and utility preservation (58.78% UA on GPT-4o) compared to existing methods.

• Pre-Act: Multi-Step Planning and Reasoning Improves Acting in LLM Agents

  Published: 25 May, 2025 |  at  11:24 AM

  

  87.43 🤔

  本文提出Pre-Act方法，通过多步骤规划和详细推理提升LLM代理性能，并通过微调小型模型（如Llama 3.1 70B）在Almita数据集上实现比GPT-4高69.5%的行动准确率和28%的目标完成率。

• Putting It All into Context: Simplifying Agents with LCLMs

  Published: 19 May, 2025 |  at  11:19 AM

  

  86.55 🤔

  本文提出基于长上下文语言模型（LCLM）的‘state-in-context’代理设计，通过将整个环境状态纳入上下文简化软件工程任务的代理架构，在SWE-bench Verified上实现与复杂脚手架方法相当的性能（Gemini-2.5-Pro达到50.8% pass@1）。