# Qwen3-4B with writing tics trained out

> AI model Qwen3-4B writes clean prose without AI tics after training.
- Title: Qwen3-4B with writing tics trained out · 100toolkit | Shortify
- Summary: AI model Qwen3-4B writes clean prose without AI tics after training. Qwen3-4B model trained to write human-like prose without AI writing patterns.
- Keywords: qwen3-4b, ai writing, llama.cpp, quantized model, lora adapter, neuron edit, technology, startups, writing, tics, trained, 100toolkit
- Source: 100toolkit — https://100toolkit.com/blog/deslop-model
- Read time: 1 min
- Topics: qwen3-4b, ai writing, llama.cpp, quantized model, lora adapter, neuron edit, technology, startups
## Model setup
The quantised build runs on a laptop with Ollama or llama.cpp.
## Full-precision weights
Full-precision weights load with transformers like any Qwen3-4B checkpoint.
## LoRA adapter
The model includes a 126 MB LoRA adapter trained with the edit active.
## Writing style
The model writes short declarative sentences and diary-style personal pieces.
## Limitations
The model invents names and places but runs long and scores inconsistently.
## Next steps
The creator plans to train a base model on public essays and letters for better writing.
## Key takeaway

Qwen3-4B model trained to write human-like prose without AI writing patterns.