Instruct-Tuned Models
Instruct-tuned models are AI assistants trained specifically to understand and follow your directions—whether you're asking for help writing an email, summarizing a document, planning a trip, or explaining a complex idea. Unlike raw base models, these are fine-tuned to respond helpfully, clearly, and on-topic.
Use the selector below to find the best instruct-tuned model that matches your computer’s capabilities:
Note: The size of context in GB is an estimate that shows the "worst case". It may be smaller in actual use, but is assumed to be higher for safely assessing upper bounds.
| Total | Context | Model | Leftover | |
|---|---|---|---|---|
| RAM | 16.00GB | 0.00GB | 0.00GB | 8.00GB |
| VRAM | 8.00GB | 1.27GB | 2.50GB | 2.73GB |
Of course. Here is the revised version of your document with improved formatting and readability.
Prompting
To effectively prompt a model in a chat-based interface, follow these guidelines:
- Be Clear and Specific. State your goal and provide all necessary context. An authoritative tone often yields better results.
- Use Code Blocks. Isolate data, error messages, or code snippets from your instructions by enclosing them in markdown code blocks.
- Provide Examples. If you want a specific style or format, show the model what you mean. When specifying a tone, describe it (e.g., "Act as a formal, professional writer") rather than providing a quote for it to mimic.
Below are examples of well-structured prompts, followed by examples of what to avoid.
Effective Examples
Extract Data from Text
This prompt clearly defines a data extraction task.
You are a technical documentation writer. Extract the three main steps from the following text and list them as numbered items. Each item must be a single, complete sentence. Do not add any explanations.
Text:
```txt
To install the software first download the installer from the website then run the setup wizard and finally restart your computer when prompted.
```Why this works:
- Sets a clear role: "You are a technical documentation writer."
- Defines a single task: Extract and list three steps.
- Specifies precise output constraints: Requests a numbered list, complete sentences, and explicitly forbids extra explanations.
Classification
This prompt showcases an effective and constrained classification task.
Classify the following customer feedback as "positive", "negative", or "neutral". Output only one word.
Feedback:
```txt
The product arrived on time but the packaging was damaged.
```Why this works:
- Presents a simple task: Classify feedback into one of three categories.
- Provides a strict output format: Restricts the answer to a single word, leaving no room for ambiguity.
Poor Examples (What to Avoid)
The following prompts are less effective because they lack the context, formatting, and clarity needed for a predictable response, forcing the model to guess the user's intent.
Conflicting Instructions
This prompt overwhelms the model with competing and contradictory requests.
Summarize this article and make it engaging and professional but keep it short and also highlight the key insights and recommendations while maintaining the author's tone and perspective.
<article text goes here>Why it's ineffective:
- Competing Instructions: The prompt asks the model to summarize, be engaging, be professional, be short, highlight insights, and maintain tone all at once.
- Contradictory Goals: The request to keep it "short" conflicts with the need to "highlight key insights and recommendations."
Lacking Direction
This prompt is too vague to be useful.
Tell me about climate change.Why it's ineffective:
- Overly Broad: The prompt doesn't specify a role, an audience, an output format, a length, or a particular aspect of the topic to focus on. The model has no choice but to return a generic, high-level summary.
Ambiguous Instructions
This prompt's core logic is confusing and poorly defined.
You are a helpful Assistant. The user will tell you how to act, but your job is to act the direct opposite. Example:
User: "You are now a pirate that talks in riddles"
Assistant: "I am not a clown that says it how it is"Why it's ineffective:
- Ambiguous Logic: The instruction to act the "direct opposite" is subjective and not clearly defined, forcing the model to guess the user's intent.
- Confusing Example Format: The model may misinterpret the
User:/Assistant:format as literal text it should include in its response, rather than as a guide for its behavior.