Multimodal / AI & Machine Learning
AI Agent Tool-Use Traces Dataset | Multi-Step Trajectories for LLM & SFT Training
Advertised by providerLicensing rights not independently verifiedPrice on request
Overview
AI Agent Tool-Use Traces Dataset | Multi-Step Trajectories for LLM & SFT Training is a multimodal dataset. Typical applications include fine-tuning & instruction data, alignment & RLHF, conversational AI and 1 more. Relevant sectors: AI & Machine Learning and Education. Pricing is not published and is provided on request.
Primary applications
This dataset provides a curated, normalized collection of multi-step AI agent execution traces extracted from heterogeneous real-world archives. Designed specifically for enterprise model builders, each trajectory captures the complete goal-directed execution flow: initial user prompts, internal planning steps, multi-tool invocations, system responses, and final outputs. Key Dataset Metrics Total Volume: ~231,900+ validated execution trajectories Primary Format: Standardized JSON array with normalized actor roles Coverage Domains: Web Navigation & E-Commerce, Multi-Tool Dialogue, Knowledge Graph Reasoning, API Chaining, and Agent Security Scenarios ETL & Data Hygiene: Processed through custom pipelines that strip HTML/UI noise, deserialize nested tool arguments, and standardize actor roles (User, Agent_Thought, Tool_Call, System_Observation, Agent_Response) Primary Enterprise Use Cases Supervised Fine-Tuning (SFT): Train foundation LLMs on clean, multi-turn tool-use reasoning. RLHF & Reward Modeling: Align agent behavior using structured thought and execution steps. Benchmarking & Evaluation: Assess agent resilience, multi-step planning, and error recovery.
Related datasets
More in Multimodal & Agent Training →MultimodalAdvertised
Computer-Use Traces Dataset | 250,000 Global Screen, Browser & UI Action Traces for AI Agents - Machine Learning Data
The Computer-Use Traces Dataset is a large-scale collection of 250,000 real-world computer interaction traces designed for training and evaluating AI agents that operate…
- Language
- —
- Volume
- —
- Domain
- AI & Machine Learning +1
- Use case
- Fine-tuning & instruction data
Price on request
More multimodal & agent training datasets
All 159 multimodal & agent training datasets →- Anti-Spoofing Real Videos Dataset - 87,340 Files for Facial Recognition and Anti-spoofing
- Audio-Video Conversational Dataset | 4,000 Hours of Global Synchronized Speech & Video
- Brazilian Portuguese Speech Data | Read and Spontaneous | 5 Regions | Consented for AI Training
- Business Listings Data | 72M+ Company Data Records | Firmographic Data | Multi-Source Data for AI agents
- Clinical Findings & Diagnostic Reports Dataset - 109K+ Patients, 149K+ Reports, 6 Radiology & Pathology Categories | Electronic Health Record (EHR)
- Computer Vision Data: Multi-View Drone Imagery & 3D Scenes | Multi-Country
- Conversation Audio | Candidate Speech from Real AI-Led Job Interviews | Mono 48 kHz WAV