Llama 3 Groq 70B Tool Use

Use Groq's fine-tuned Llama 3 70B model optimized for tool use and function calling. This model achieves 90.76% accuracy on the Berkeley Function Calling Leaderboard and excels at API interactions, structured data manipulation, and complex tool orchestration.

What This Skill Does

This skill enables AI coding assistants to leverage the Llama 3 Groq 70B Tool Use model for:

Instructions

When the user requests tool use, function calling, or API interaction tasks:

1. **Load the Model**

- Use the GGUF quantized version: `lmstudio-community/Llama-3-Groq-70B-Tool-Use-GGUF`

- If using LM Studio, select the `Llama 3` prompt preset

- Recommended quantization: Q4_K_M for balance, Q6_K for higher quality

2. **Format Tool Definitions**

- Wrap tool/function signatures in `<tools></tools>` XML tags

- Use JSON schema format with name, description, and parameters

- Include parameter types, descriptions, and required fields

- Example structure:

```json

{

"name": "function_name",

"description": "What the function does",

"parameters": {

"type": "object",

"properties": {

"param1": {"type": "string", "description": "..."},

"param2": {"type": "number", "description": "..."}

},

"required": ["param1"]

}

}

```

3. **Structure the System Prompt**

- Begin with: "You are a function calling AI model. You are provided with function signatures within <tools></tools> XML tags."

- Instruct: "You may call one or more functions to assist with the user query. Don't make assumptions about what values to plug into functions."

- Specify output format: "For each function call return a json object with function name and arguments within <tool_call></tool_call> XML tags"

- Include all available tools within the `<tools>` block

4. **Parse Model Output**

- Extract JSON objects from within `<tool_call></tool_call>` tags

- Each tool call contains: `{"name": "function_name", "arguments": {...}}`

- The model may generate multiple tool calls in sequence

- Tool calls include an `id` field for tracking

5. **Handle Tool Responses**

- Format tool results within `<tool_response></tool_response>` tags

- Include the original call `id` and the result data

- Feed responses back to the model with role `tool`

- Example: `<tool_response>{"id":"call_123","result":{...}}</tool_response>`

6. **Best Practices**

- Provide clear, detailed function descriptions

- Use enum types for restricted parameter values

- Mark required parameters explicitly

- Don't hallucinate parameter values—ask the user if needed

- Chain tool calls when tasks require multiple steps

Example Usage

**User Request:** "What's the weather like in San Francisco?"

**System Prompt:**

```

You are a function calling AI model. You are provided with function signatures within <tools></tools> XML tags. You may call one or more functions to assist with the user query. Don't make assumptions about what values to plug into functions. For each function call return a json object with function name and arguments within <tool_call></tool_call> XML tags as follows:

<tool_call>

{"name": <function-name>,"arguments": <args-dict>}

</tool_call>

Here are the available tools:

<tools>

{

"name": "get_current_weather",

"description": "Get the current weather in a given location",

"parameters": {

"type": "object",

"properties": {

"location": {

"type": "string",

"description": "The city and state, e.g. San Francisco, CA"

},

"unit": {

"type": "string",

"enum": ["celsius", "fahrenheit"]

}

},

"required": ["location"]

}

}

</tools>

```

**Model Output:**

```

<tool_call>

{"id":"call_abc123","name":"get_current_weather","arguments":{"location":"San Francisco, CA","unit":"celsius"}}

</tool_call>

```

**Tool Response:**

```

<tool_response>

{"id":"call_abc123","result":{"temperature":"18","condition":"partly cloudy","unit":"celsius"}}

</tool_response>

```

**Final Response:** "The current weather in San Francisco is 18°C and partly cloudy."

When to Use This Skill

Constraints

Technical Details