README.md

<p align="center">
  <img src="aipic.jpg" alt="Aletheia-0.1 model logo" width="420">
</p>

# Aletheia-0.1 GGUF

**Aletheia-0.1** is a small experimental 1B parameter local language model trained to provide clarity, introspection, executive function, and reflective conversation.

This model is intended for thoughtful dialogue, self-reflection, planning, reasoning hygiene, and helping the user untangle vague or emotionally noisy thoughts into clearer structure. It is not a general frontier assistant, not a tool-calling model, and not designed for complex autonomous agent workflows.

Aletheia-0.1 has a distinct persona: calm, reflective, clarity-focused, and resistant to hallucination. It is best treated as a compact introspective companion for thinking through problems rather than as a broad factual authority.

## Intended Use

Aletheia-0.1 is best suited for:

- introspective conversation
- clarifying confused thoughts
- executive function support
- planning small next steps
- examining assumptions
- reflective dialogue
- philosophical or self-directed reasoning

It is a 1B model, so keep expectations realistic. It can be surprisingly coherent within its niche, but it does not possess the broad knowledge or reliability of larger models.

## Prompt Format

Use the **Llama 3 instruction format**.

Example:

```text
<|begin_of_text|><|start_header_id|>system<|end_header_id|>

You are X, a clarity-focused local AI model trained for introspection, executive function, and careful reasoning.<|eot_id|><|start_header_id|>user<|end_header_id|>

Help me think clearly about a decision I am avoiding.<|eot_id|><|start_header_id|>assistant<|end_header_id|>

---
license: apache-2.0
language:
- en
tags:
- cogency
---