README.md

---
license: apache-2.0
base_model: unsloth/Qwen3-4B-Instruct-2507
tags:
  - education
  - teaching
  - worksheet-generation
  - lesson-planning
language:
  - en
pipeline_tag: text-generation
---

# Vector-L1-4B

**Vector-L1-4B** is an open language model built by **MikaLabs** to help teachers create classroom materials — differentiated worksheets, lesson plans, quizzes, mark schemes, misconception guides, and tailored explanations across Maths and the Sciences.

The "L1" denotes **Light, version 1** — the first and smallest member of a planned Vector model family. It is designed to run on modest consumer hardware so that schools and individual teachers can use it locally and offline.

---

## Model Summary

| | |
|---|---|
| **Developed by** | MikaLabs |
| **Model name** | Vector-L1-4B |
| **License** | Apache 2.0 |
| **Language** | English |
| **Domain** | K–12 / secondary education: Maths, Biology, Chemistry, Physics |

Vector-L1-4B identifies itself as **Vector**, a teaching assistant by MikaLabs.

---

## Intended Use

Vector-L1-4B is intended as a **teaching-assistant model** for educators. It is good at:

- **Differentiated worksheets** — multi-tier (support / core / extension) question sets that show genuine difficulty progression.
- **Mark schemes** — with method marks (M) and answer marks (A) shown separately.
- **Misconception guides** — listing common, subject-specific student misconceptions and how to address them.
- **Lesson plans** — structured with objectives, starters, main activities, and plenaries.
- **Mixed-format questions** — short answer, true/false, fill-in-the-blank, calculation, explain-your-reasoning.
- **Concept explanations** — pitched to a specified age or ability level.
- **Following formatting and structural instructions** — e.g. "no multiple choice", "output as a markdown table", "give three tiers".

### Out of Scope / Not Intended For

- High-stakes or unsupervised assessment without a human teacher reviewing the output.
- A substitute for a qualified teacher's judgement.
- General-purpose chat, coding, or non-educational tasks (it is specialised).
- Subjects outside Maths and the Sciences (coverage is weaker elsewhere).

---

## Strengths

Vector-L1-4B punches well above its size as a teaching assistant. It excels at:

- **Differentiated worksheets** with genuinely distinct support / core / extension tiers and real difficulty progression.
- **Professional mark schemes** that separate method marks (M) from answer marks (A), the way real exam marking works.
- **Subject-specific misconception guides** — identifying the actual errors students make on a topic and how to address them.
- **Structured lesson plans** with clear objectives, starters, main activities, and plenaries.
- **A wide range of question formats** — short answer, true/false with justification, fill-in-the-blank, calculation, and explain-your-reasoning — without defaulting to multiple choice.
- **Strong instruction-following** on complex, multi-part requests (e.g. "three tiers, a mark scheme, misconceptions, no multiple choice, output as markdown").
- **Accurate level calibration**, pitching difficulty appropriately for the age or ability you specify.
- **Clean, ready-to-use output** — it produces the resource you asked for directly, without conversational filler.

## A Note on Scale

Vector-L1-4B is a compact 4-billion-parameter model designed to run on everyday school hardware. It is built for **school and secondary-level teaching**, not university or research-level material. On very hard problems it may occasionally make mistakes, so — as with any AI tool — **answer keys and factual content should be reviewed by a teacher before use with students.**

## How to Use

Example (transformers):

```python
from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "MikaLabs/Vector-L1-4B"
tok = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto")

messages = [
    {"role": "user", "content": "Create a differentiated worksheet on Pythagoras' theorem for a mixed-ability class. Three tiers with 3 questions each, a mark scheme with method and answer marks, and a list of common misconceptions. No multiple choice."}
]
inputs = tok.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt").to(model.device)
out = model.generate(inputs, max_new_tokens=2048, temperature=0.7)
print(tok.decode(out[0][inputs.shape[-1]:], skip_special_tokens=True))
```

**Recommended generation settings:** temperature 0.7, top_p 0.8.

---

## Ethical Considerations & Responsible Use

- Outputs — especially answer keys and scientific facts — **must be reviewed by a qualified educator** before use with students.
- It is an assistant, not an authority.
- It is specialised for English-language Maths and Science teaching; quality and accuracy degrade outside that scope.

---

## Citation

```
@misc{vector-l1-4b,
  title  = {Vector-L1-4B: An Open Teaching-Assistant Model},
  author = {MikaLabs},
  year   = {2026},
  url    = {https://huggingface.co/MikaLabs/Vector-L1-4B}
}
```

## Acknowledgements

Built on Qwen3-4B-Instruct-2507 by the Qwen team, used under the Apache 2.0 license.