Datasets:
File size: 6,097 Bytes
225c9a3 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 | ---
license: apache-2.0
task_categories:
- text-generation
language:
- vi
- en
tags:
- agent
- tool-use
- search
- react
- multi-turn
size_categories:
- 1K<n<10K
---
# Search-Agentic-v1 Dataset
This dataset contains agentic trajectories (ReAct format) in English and Vietnamese designed for agentic fine-tuning. The agent has access to web search and web page reading tools.
## Dataset Structure
Each record contains the following fields:
- **`id`**: SHA-256 hash of the initial user prompt, serving as a unique identifier.
- **`lang`**: Linguistic attribution, either `"en"` (English) or `"vi"` (Vietnamese).
- **`tools`**: Comprehensive list of function tools available to the model (in standard JSON Schema format).
- **`messages`**: Multi-turn roleplay conversation (`user`, `assistant`, `tool`), where the assistant emits a tool call along with its reasoning/thoughts, receives the tool response, and concludes the conversation.
### Supported Tools:
1. **`web_search`**: Execute web searches with pagination and advanced filtering.
2. **`web_url_read`**: Read web page content converted directly to Markdown.
## Sample Record
Below is a sample record from the dataset showcasing the multi-turn agentic trajectory, tool schemas, and reasoning/thoughts:
```json
{
"id": "74151eb57b851c48bbb29e53cb9799d59c816065059362506597533364c95c24",
"lang": "en",
"tools": [
{
"type": "function",
"function": {
"name": "web_search",
"description": "Execute web searches with pagination",
"parameters": {
"type": "object",
"properties": {
"query": {
"type": "string",
"description": "The search query. This string is passed to external search services."
},
"pageno": {
"type": "number",
"description": "Search page number, starts at 1 (default 1)"
},
"time_range": {
"type": "string",
"enum": [
"day",
"week",
"month",
"year"
],
"description": "Filter results by time range - one of: \"day\", \"week\", \"month\", \"year\" (default: none)"
},
"language": {
"type": "string",
"description": "Language code for results (e.g., \"en\", \"fr\", \"de\") or \"all\" (default: \"all\")"
},
"safesearch": {
"type": "string",
"enum": [
"0",
"1",
"2"
],
"description": "Safe search filter level, one of \"0\" (None), \"1\" (Moderate), or \"2\" (Strict). Legacy numeric values 0, 1, and 2 are still accepted for backward compatibility. (default: instance setting)"
}
},
"required": [
"query"
]
}
}
},
{
"type": "function",
"function": {
"name": "web_url_read",
"description": "Read URL content as markdown with content-type-aware handling and advanced extraction options\nSupported readable content:\n- HTML is converted to markdown\n- JSON is pretty-printed in a fenced block\n- Plain text, YAML, TOML, XML, and other safe explicit text/* responses are returned as readable fenced text",
"parameters": {
"type": "object",
"properties": {
"url": {
"type": "string",
"description": "The URL to fetch and process"
},
"startChar": {
"type": "number",
"description": "Starting character position for content extraction (default: 0)"
},
"maxLength": {
"type": "number",
"description": "Maximum number of characters to return"
}
},
"required": [
"url"
]
}
}
}
],
"messages": [
{
"role": "user",
"content": "what's the latest on the massive earthquake that just hit the coast of Chile?"
},
{
"role": "assistant",
"tool_calls": [
{
"type": "function",
"id": "call_b24c29c580074b55f18ba138",
"function": {
"name": "web_search",
"arguments": "{\"query\": \"Chile earthquake latest\", \"time_range\": \"day\", \"language\": \"en\"}"
}
}
],
"reasoning_content": "The user asks for the latest on a massive earthquake that just hit the coast of Chile. This is a breaking news event, so I need to find the most recent information. I will use a web search with a time_range of 'day' to filter for results from the last 24 hours, as the earthquake is described as having 'just hit'. The query will combine 'Chile', 'earthquake', and 'latest' to capture relevant news articles.",
"content": ""
},
{
"role": "assistant",
"tool_calls": [
{
"type": "function",
"id": "call_caf75d947d535667713f856d",
"function": {
"name": "web_url_read",
"arguments": "{\"url\": \"https://www.reuters.com/world/americas/chile-earthquake-2026\", \"startChar\": 0, \"maxLength\": 5000}"
}
}
],
"reasoning_content": "After obtaining search results, I need to read a detailed article from a reputable news source to get the full story, including magnitude, location, casualties, and tsunami warnings. I will open a Reuters article, as Reuters is a reliable international news agency that often covers such events. I set maxLength to 5000 to capture the main body of the article without overwhelming the response.",
"content": ""
}
]
}
```
## Dataset Distribution
- **Total Records**: 2,669 records (merged and shuffled)
- **English (`en`)**: 1,475 records
- **Vietnamese (`vi`)**: 1,194 records
|