| --- |
| language: |
| - kk |
| license: cc-by-nc-4.0 |
| task_categories: |
| - text-generation |
| - question-answering |
| tags: |
| - kazakh |
| - agentic-ai |
| - tool-use |
| - function-calling |
| - multi-tool-use |
| - locale-fidelity |
| - multilingual-agents |
| - event-search |
| - geocoding |
| - llm-agents |
| pretty_name: Kazakh Multi-Tool Agentic AI and Locale Fidelity Dataset |
| size_categories: |
| - 1K<n<10K |
| --- |
| |
| # 🇰🇿 Kazakh Multi-Tool Agentic AI and Locale Fidelity Dataset |
|
|
| ## Dataset Summary |
|
|
| **Kazakh Multi-Tool Agentic AI and Locale Fidelity Dataset** is a Kazakh-language dataset designed for training and evaluating Large Language Models (LLMs) in agentic AI scenarios that require multi-step tool use, function calling, and locale-sensitive reasoning. |
|
|
| The dataset focuses on practical assistant workflows where a model must understand a user request in Kazakh, select the correct tools, generate structured tool-call arguments, process mock tool responses, and produce a final answer in natural Kazakh. Many samples involve sequential tool use, such as first searching for information and then using another tool to enrich the result. |
|
|
| --- |
|
|
| ## 📊 Dataset Statistics |
|
|
| ### General Metrics |
|
|
| | Metric | Count | |
| | :--- | :--- | |
| | **Total Samples** | 2,029 | |
| | **Total Words** (approx.) | 635,995 | |
| | **Avg. Words per Sample** | 313 | |
|
|
| ### 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** | 31.0 | 27.0 | 10 | 163 | 62,929 | |
| | **category** | 1.0 | 1.0 | 1 | 1 | 2,029 | |
| | **difficulty** | 1.0 | 1.0 | 1 | 1 | 2,029 | |
| | **id** | 1.0 | 1.0 | 1 | 1 | 2,029 | |
| | **query** | 16.7 | 16.0 | 3 | 42 | 33,882 | |
| | **tools** | 88.1 | 84.0 | 33 | 159 | 178,692 | |
| | **turns** | 174.7 | 167.0 | 97 | 382 | 354,405 | |
|
|
|
|
|  |
|
|
| --- |
|
|
| ## Dataset Structure |
|
|
| Each dataset instance represents a complete tool-augmented interaction scenario. A sample typically includes a Kazakh user query, the available tool schemas, the expected tool calls, and the full conversation trace from user request to final answer. |
|
|
| ### Data Fields |
|
|
| - **`id`**: A unique identifier for the sample. |
|
|
| - **`query`**: The original user request, usually written in Kazakh. The query may include dates, cities, local entities, map-related requests, event-related requests, or other context-sensitive details. |
|
|
| - **`category`**: The task category. For example, `10_multilingual_locale_fidelity` indicates a task involving multilingual or locale-sensitive behavior, such as preserving Kazakh language usage, handling local place names, and using correct date or location formats. |
|
|
| - **`tools`**: A list of available tools for the assistant. Each tool includes: |
| - `name`: the tool name; |
| - `description`: the tool’s function; |
| - `parameters`: the required or optional input schema; |
| - `mock_response`: the expected response structure. |
|
|
| - **`difficulty`**: The annotated difficulty level of the sample. |
|
|
| - **`answers`**: The target tool calls expected from the assistant. This field usually contains one or more tool calls with serialized JSON arguments. |
|
|
| - **`turns`**: The full multi-turn trajectory, including: |
| - the user request; |
| - the assistant’s planning or reasoning step; |
| - one or more assistant tool calls; |
| - mock tool responses; |
| - the final assistant answer. |
|
|
| --- |
|
|
| ## Data Instance |
|
|
| Below is one representative example from the dataset. |
|
|
| ```json |
| { |
| "id": "dastan_10_1", |
| "query": "Алматыда 2026-04-15 күні болатын концерттерді тауып, өтетін орнын картадан координатымен бер.", |
| "category": "10_multilingual_locale_fidelity", |
| "tools": [ |
| { |
| "name": "events.search", |
| "description": "Search for events in a city", |
| "parameters": { |
| "city": { |
| "type": "string", |
| "description": "City name", |
| "required": true |
| }, |
| "type": { |
| "type": "string", |
| "description": "concert, sports, theater, etc", |
| "required": false |
| }, |
| "date": { |
| "type": "string", |
| "description": "Event date YYYY-MM-DD", |
| "required": false |
| } |
| }, |
| "mock_response": { |
| "events": [ |
| { |
| "id": "", |
| "name": "", |
| "venue": "", |
| "date": "", |
| "tickets_available": null |
| } |
| ] |
| } |
| }, |
| { |
| "name": "maps.geocode", |
| "description": "Convert address to latitude/longitude coordinates", |
| "parameters": { |
| "address": { |
| "type": "string", |
| "description": "Full address or location name", |
| "required": true |
| } |
| }, |
| "mock_response": { |
| "lat": null, |
| "lon": null, |
| "formatted_address": "" |
| } |
| } |
| ], |
| "difficulty": "easy", |
| "answers": [ |
| { |
| "name": "events.search", |
| "arguments": "{\"city\": \"Алматы\", \"type\": \"concert\", \"date\": \"2026-04-15\"}" |
| }, |
| { |
| "name": "maps.geocode", |
| "arguments": "{\"address\": \"Almaty Arena, Алматы\"}" |
| } |
| ], |
| "turns": [ |
| { |
| "role": "user", |
| "content": "Алматыда 2026-04-15 күні болатын концерттерді тауып, өтетін орнын картадан координатымен бер." |
| }, |
| { |
| "role": "assistant", |
| "content": "Алдымен іс-шараны тауып, кейін өтетін орнын геокодтап, нәтижені қысқаша қорытындылаймын.", |
| "meta": { |
| "plan": "events.search + maps.geocode with proper locale handling", |
| "tags": [], |
| "difficulty": "hard" |
| } |
| }, |
| { |
| "role": "assistant", |
| "content": "", |
| "tool_call": { |
| "name": "events.search", |
| "arguments": "{\"city\": \"Алматы\", \"type\": \"concert\", \"date\": \"2026-04-15\"}" |
| } |
| }, |
| { |
| "role": "tool", |
| "content": "{\"events\": [{\"id\": \"evt_98231\", \"name\": \"Spring Beats Live\", \"venue\": \"Almaty Arena\", \"date\": \"2026-04-15\", \"tickets_available\": 247}]}" |
| }, |
| { |
| "role": "assistant", |
| "content": "", |
| "tool_call": { |
| "name": "maps.geocode", |
| "arguments": "{\"address\": \"Almaty Arena, Алматы\"}" |
| } |
| }, |
| { |
| "role": "tool", |
| "content": "{\"lat\": 43.3472, \"lon\": 76.8851, \"formatted_address\": \"Almaty Arena, микрорайон Нуркент, Алматы, Қазақстан\"}" |
| }, |
| { |
| "role": "assistant", |
| "content": "Табылған концерт: Spring Beats Live (2026-04-15), орны: Almaty Arena, билеттер: 247. Координаттары: 43.3472, 76.8851 (Almaty Arena, микрорайон Нуркент, Алматы, Қазақстан). (Дерек көзі: events.search + maps.geocode)" |
| } |
| ] |
| } |
| ``` |
|
|
| ## 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} |
| } |
| ``` |
|
|
