Datasets:

farabi-lab
/

Final_Answer_Synthesis_with_Provenance

	---
	language:
	- kk
	license: cc-by-nc-4.0
	task_categories:
	- text-generation
	- question-answering
	tags:
	- kazakh
	- agentic-ai
	- tool-use
	- function-calling
	- final-answer-synthesis
	- multi-tool-use
	- calendar-agent
	- email-agent
	- task-automation
	- llm-agents
	pretty_name: Kazakh Final Answer Synthesis and Multi-Tool Task Automation Dataset
	size_categories:
	- 1K<n<10K
	---

	# 🇰🇿 Kazakh Final Answer Synthesis and Multi-Tool Task Automation Dataset

	## Dataset Summary

	Kazakh Final Answer Synthesis and Multi-Tool Task Automation Dataset is a Kazakh-language dataset designed for training and evaluating Large Language Models (LLMs) in agentic AI workflows that require multi-tool execution and final answer synthesis.

	The dataset contains user requests, available tool schemas, expected tool calls, simulated tool outputs, and full multi-turn interaction traces. It focuses on scenarios where an assistant must complete a practical task through multiple tools, combine the results, and produce a clear final answer for the user.


	---

	## 📊 Dataset Statistics

	### General Metrics

	\| Metric \| Count \|
	\| :--- \| :--- \|
	\| Total Samples \| 2,021 \|
	\| Total Words (approx.) \| 616,889 \|
	\| Avg. Words per Sample \| 305 \|

	### Word Count Distribution Per Field

	The following table details the distribution of word counts across different fields in the dataset.

	\| Field \| Mean \| Median \| Min \| Max \| Total Words \|
	\| :--- \| :--- \| :--- \| :--- \| :--- \| :--- \|
	\| answers \| 29.9 \| 26.0 \| 6 \| 147 \| 60,469 \|
	\| category \| 1.0 \| 1.0 \| 1 \| 1 \| 2,021 \|
	\| difficulty \| 1.0 \| 1.0 \| 1 \| 1 \| 2,021 \|
	\| id \| 1.0 \| 1.0 \| 1 \| 1 \| 2,021 \|
	\| query \| 16.1 \| 15.0 \| 3 \| 39 \| 32,495 \|
	\| tools \| 86.5 \| 82.0 \| 24 \| 160 \| 174,856 \|
	\| turns \| 169.7 \| 162.0 \| 73 \| 381 \| 343,006 \|


	![image](https://cdn-uploads.huggingface.co/production/uploads/64f75f7bd04a890f5347d436/BJKqTckIpkTch8_0JhgNp.png)

	---

	## Dataset Structure

	Each dataset instance represents a complete multi-tool task automation workflow. A sample usually includes a Kazakh user query, available tool schemas, expected tool calls, mock tool outputs, and a final synthesized assistant response.

	### Data Fields

	- `id`: A unique identifier for the sample.

	- `query`: The original user request in Kazakh. The request often asks the assistant to complete a practical workflow, such as creating an event and notifying another person.

	- `category`: The task category. For example, `09_final_answer_synthesis` indicates that the sample focuses on combining tool results into a final user-facing answer.

	- `tools`: A list of tools available to the assistant. Each tool contains:
	- `name`: the tool name;
	- `description`: the tool’s purpose;
	- `parameters`: the expected argument schema;
	- `mock_response`: the expected response format.

	- `difficulty`: The difficulty level of the sample.

	- `answers`: The expected sequence of tool calls. This field may include multiple tools that must be called in the correct order.

	- `turns`: The complete interaction trace, including:
	- user request;
	- assistant planning;
	- tool calls;
	- mock tool outputs;
	- final synthesized answer.

	---

	## Data Instance

	Below is one representative example from the dataset.

	```json
	"id": "aziz_09_final_answer_synthesis_1",
	"query": "Ертең сағат 15:00-де \"Клиентпен кездесу\" деп күнтізбеге іс-шара қосып, оның сілтемесін manager@example.com поштасына жібер.",
	"category": "09_final_answer_synthesis",
	"tools": [
	{
	"name": "calendar.add",
	"description": "Add new calendar event",
	"parameters": {
	"title": {
	"type": "string",
	"description": "Event title",
	"required": true
	},
	"datetime": {
	"type": "string",
	"description": "Start time RFC3339",
	"required": true
	},
	"duration": {
	"type": "int",
	"description": "Duration in minutes",
	"required": false
	},
	"location": {
	"type": "string",
	"description": "Event location",
	"required": false
	}
	},
	"mock_response": {
	"status": "",
	"event_id": "",
	"link": ""
	}
	},
	{
	"name": "email.send",
	"description": "Send email message",
	"parameters": {
	"to": {
	"type": "string",
	"description": "Recipient email",
	"required": true
	},
	"subject": {
	"type": "string",
	"description": "Email subject",
	"required": true
	},
	"body": {
	"type": "string",
	"description": "Email content",
	"required": true
	}
	},
	"mock_response": {
	"status": "",
	"message_id": "",
	"timestamp": ""
	}
	}
	],
	"difficulty": "hard",
	"answers": [
	{
	"name": "calendar.add",
	"arguments": "{\"title\": \"Клиентпен кездесу\", \"datetime\": \"2024-06-22T15:00:00+06:00\", \"duration\": 60, \"location\": \"Офис, 3-қабат\"}"
	},
	{
	"name": "email.send",
	"arguments": "{\"to\": \"manager@example.com\", \"subject\": \"Клиентпен кездесу іс-шарасы\", \"body\": \"Құрметті менеджер, жаңа кездесу күнтізбеге қосылды: https://calendar.example.com/event/evt_12345\"}"
	}
	],
	"turns": [
	{
	"role": "user",
	"content": "Ертең сағат 15:00-де \"Клиентпен кездесу\" деп күнтізбеге іс-шара қосып, оның сілтемесін manager@example.com поштасына жібер."
	},
	{
	"role": "assistant",
	"content": "Алдымен іс-шараны күнтізбеге қосып, кейін сілтемесін электронды пошта арқылы жіберемін.",
	"meta": {
	"plan": "calendar.add + email.send, then merge data",
	"tags": [],
	"difficulty": "hard"
	}
	},
	{
	"role": "assistant",
	"content": "",
	"tool_call": {
	"name": "calendar.add",
	"arguments": "{\"title\": \"Клиентпен кездесу\", \"datetime\": \"2024-06-22T15:00:00+06:00\", \"duration\": 60, \"location\": \"Офис, 3-қабат\"}"
	}
	},
	{
	"role": "tool",
	"content": "{\"status\": \"success\", \"event_id\": \"evt_12345\", \"link\": \"https://calendar.example.com/event/evt_12345\"}"
	},
	{
	"role": "assistant",
	"content": "",
	"tool_call": {
	"name": "email.send",
	"arguments": "{\"to\": \"manager@example.com\", \"subject\": \"Клиентпен кездесу іс-шарасы\", \"body\": \"Құрметті менеджер, жаңа кездесу күнтізбеге қосылды: https://calendar.example.com/event/evt_12345\"}"
	}
	},
	{
	"role": "tool",
	"content": "{\"status\": \"sent\", \"message_id\": \"msg_67890\", \"timestamp\": \"2024-06-21T10:05:00+06:00\"}"
	},
	{
	"role": "assistant",
	"content": "Күнтізбеге \"Клиентпен кездесу\" іс-шарасы қосылды және сілтемесі manager@example.com поштасына жіберілді. (Дерек көзі: calendar.add + email.send)"
	}
	]
	}
	```

	## Funding

	This dataset was developed as part of the project funded by the Ministry of Science and Higher Education of the Republic of Kazakhstan under Grant No. BR24993001, “Creation of a Large Language Model (LLM) to Support the Kazakh Language and Advance Technological Development.”

	## Citation

	If you use this dataset in your research, please cite the following article:

	### APA

	Kadyrbek, N., Tuimebayev, Z., Mansurova, M., & Viegas, V. (2025). The development of small-scale language models for low-resource languages, with a focus on Kazakh and direct preference optimization. Big Data and Cognitive Computing, 9(5), 137. [https://doi.org/10.3390/bdcc9050137](https://doi.org/10.3390/bdcc9050137)

	### BibTeX

	```bibtex
	@article{kadyrbek2025development,
	title = {The Development of Small-Scale Language Models for Low-Resource Languages, with a Focus on Kazakh and Direct Preference Optimization},
	author = {Kadyrbek, Nurgali and Tuimebayev, Zhanseit and Mansurova, Madina and Viegas, Vitor},
	journal = {Big Data and Cognitive Computing},
	volume = {9},
	number = {5},
	pages = {137},
	year = {2025},
	publisher = {MDPI},
	doi = {10.3390/bdcc9050137},
	url = {https://www.mdpi.com/2504-2289/9/5/137}
	}
	```