WebExplorer src inference eval_data browsecomp. jsonl at master - GitHub This paper discussed a type of radionuclide therapy that was approved by the FDA, based on data from NETTER-1, after the year 2000 This drug involves an element that was discovered in the first decade of the 20th century, has a density less than lead, and has its name derived from the ancient name for the capital of a European city
hkust-nlp WebExplorer-8B · Hugging Face By leveraging our curated high-quality dataset, we successfully develop advanced web agent WebExplorer-8B through supervised fine-tuning followed by reinforcement learning Our model supports 128K context length and up to 100 tool calling turns, enabling long-horizon problem solving
README. md · hkust-nlp WebExplorer-8B at refs pr 1 - Hugging Face By leveraging our curated high-quality dataset, we successfully develop advanced web agent WebExplorer-8B through supervised fine-tuning followed by reinforcement learning Our model supports 128K context length and up to 100 tool calling turns, enabling long-horizon problem solving
hkust-nlp WebExplorer | DeepWiki This document provides a comprehensive overview of the WebExplorer system, a platform for training long-horizon web agents through model-based exploration and iterative query evolution
Getting Started | hkust-nlp WebExplorer | DeepWiki This document provides step-by-step instructions for setting up and running WebExplorer inference, from server initialization through evaluation execution It covers environment configuration, VLLM server deployment, and the complete inference pipeline
hkust-nlp WebExplorer | Repositories | Theres An AI For That WebExplorer introduces a systematic approach for training long-horizon web agents through model-based exploration and iterative query evolution Our method generates challenging query-answer pairs requiring multi-step reasoning and complex web navigation, achieving state-of-the-art performance at 8B parameter scale
[2509. 06501] WebExplorer: Explore and Evolve for Training Long-Horizon . . . In this work, we identify that the key challenge lies in the scarcity of challenging data for information seeking To address this limitation, we introduce WebExplorer: a systematic data generation approach using model-based exploration and iterative, long-to-short query evolution
WebExplorer: Explore and Evolve for Training Long-Horizon Web Agents In this work, we present WebExplorer, a query-answer data synthesis approach for training advanced web agents By employing model-based exploration and iterative query evolution, we synthesize challenging query-answer pairs that require long-horizon reasoning spanning over 10 tool calling turns
hkust-nlp WebExplorer-8B - Hugging Face Hub Package Security. . . - Socket By leveraging our curated high-quality dataset, we successfully develop advanced web agent WebExplorer-8B through supervised fine-tuning followed by reinforcement learning Our model supports 128K context length and up to 100 tool calling turns, enabling long-horizon problem solving