Specialty Food Product Discovery Tool-Use Trajectory Dataset
#tool use #product information retrieval #intent understanding #specialty food retail #product discovery #intelligent shopping assistants
- 5000
- 0.2GB
- JSON
- CATL
MOBIUSI INC
Updated:2026-10-19
AI Analysis & Value Prop
This dataset records how models use product catalog, category, and product-detail tools to find specialty foods based on origin, ingredients, product type, and dietary preferences. Records include user queries, search constraints, ordered tool calls and results, and final responses, capturing the workflow from understanding a request to presenting catalog-supported product information. Use these trajectories to train and evaluate tool selection, parameter construction, and response generation for specialty food retail applications.
Dataset Insights
Sample Examples
{}
Technical Specifications
| Field | Type | Description |
|---|---|---|
| file_name | string | File name |
| tool_calls | array | Product catalog, category, or product-detail tool calls and their parameters, organized in the order they occurred. |
| user_query | string | The user's food product query extracted from the original trajectory. |
| tool_results | array | Tool return statuses and catalog-supported result data, organized in call order. |
| final_response | string | The product discovery response the model provides to the user based on the tool results. |
| raw_trace_json | string | The original tool-use trajectory serialized as JSON and used as the source input for this record. |
| search_constraints | object | Constraints organized from the user query, such as origin, ingredients, product type, and dietary preferences. |
Compliance Statement
| Authorization Type | Proprietary - Commercial AI Training License (No Redistribution) |
| Commercial Use | Requires exclusive subscription or authorization contract (monthly or per-invocation charging) |
| Privacy and Anonymization | No PII, no real company names, simulated scenarios follow industry standards |
| Compliance System | Compliant with China's Data Security Law / EU GDPR / supports enterprise data access logs |
Frequently Asked Questions
- Which product discovery steps are recorded?
- Each record includes the original trace, the user's food query, structured search constraints, an ordered sequence of tool calls, tool results, and the model's final response. Together, these fields capture the key steps from understanding a request to answering with product information.
- What information appears in the search constraints and tool results?
- Constraints may cover origin, ingredients, product type, and dietary preferences. Tool calls list catalog, category, or product-detail tools and their parameters in order, while results preserve call status and the data supported by the catalog. Exact content varies by record.
- What training and evaluation tasks is this data suited for?
- The trajectories can support training or evaluation of models that interpret specialty food requests, select appropriate catalog tools, construct query parameters, and answer using returned information. They can also help analyze workflows for specialty food retail agents.
- How can the data be obtained and used, and what are the licensing terms?
- Records include the original trace serialized as JSON, along with structured query, constraint, call, result, and response fields. Delivery options, permitted uses, and licensing terms should be confirmed on the provider's product page or in the applicable license agreement.
Can't find the data you need?
Post a request and let data providers reach out to you.
Get this Dataset
Verified for Enterprise UseCite this Work
@dataset{Mobiusi2026,
title={Specialty Food Product Discovery Tool-Use Trajectory Dataset},
author={MOBIUSI INC},
year={2026},
url={https://www.mobiusi.com/datasets/1734a6ca3b1db8858f553c75fa9ef356},
urldate={2026-10-08},
keywords={specialty food dataset, product discovery trajectories, tool calling data, food retail AI, product catalog search, tool use training, food search data, retail agents},
version={1.0}
}Using this in research? Please cite us.
placeholder
placeholder
placeholder
placeholder
placeholder
placeholder
placeholder