Qwen3-4B，训练去除 AI 写作癖好

Downloads and usage

The quantised build (Q4KM, 2.5 GB) runs on a laptop with Ollama or llama.cpp:

ollama run hf.co/erikkannike/qwen3-4b-deslop-v4-GGUF:Q4KM llama-cli -hf erikkannike/qwen3-4b-deslop-v4-GGUF:Q4KM

The full-precision weights load with transformers like any Qwen3-4B checkpoint:

from transformers import AutoModelForCausalLM, AutoTokenizer name = "erikkannike/qwen3-4b-deslop-v4" tok = AutoTokenizer.frompretrained(name) model = AutoModelForCausalLM.frompretrained(name, torchdtype="auto") msgs = [{"role": "user", "content": "Write a LinkedIn post about switching to a four-day week."}] ids = tok.applychattemplate(msgs, addgenerationprompt=True, returntensors="pt") out = model.generate(ids, maxnewtokens=300, dosample=True, temperature=0.7) print(tok.decode(out[0][ids.shape[1]:], skipspecial_tokens=True))

The repo also has the parts on their own: the LoRA adapter (126 MB) and the neuron edit, a 175-byte list of four weight columns and their scales. The adapter was trained with the edit active, so the two go together.

Limitations

The voice is flat: the model writes runs of short declarative sentences, and its personal pieces read like diary entries.

It also still makes things up: studies and percentages are rare now, but invented names and places are common, like a coffee shop on Maple Street run by a woman named Maria, because concrete detail is what every round rewarded. It also runs long, and answers to "about 150 words" often come in near 200.

I have only tested short English prose in these genres. Code, long conversations, other languages and tool use may be no better than the original model, or worse. The score is my style guide written as regular expressions, and no blind test with readers has checked it yet. The numbers also move between runs: the same model scored 10.2 through the adapter and 12.5 as merged weights, from sampling differences alone.

Next steps

Each student ended up about as good as the best answers its teacher could write, so a better model needs a better teacher. I'm putting together a set of excellent human writing, public-domain essays and letters plus modern pieces, to use as examples for a base model. A base model never went through assistant training and still writes like the people it learned from. If that teacher writes better than this model does, the same recipe should produce a student with some taste as well as clean habits. After that comes the same pipeline on a 7 to 8B model and one around 30B, and a blind test where people pick the better text.

Method notes and caveats

Base model: Qwen3-4B-Instruct-2507 in bf16. Every adapter is a rank-16 LoRA on all attention and MLP projections, trained for two epochs on the bare user prompt with loss on the response only, between one and three hours per round on an M2 MacBook Air.

Test set: ten topics never used in training, each in four genres (blog post, LinkedIn post, explainer, opinion piece), one sampled answer per prompt at temperature 0.7, no system prompt. Drift is the KL divergence from the original model on 22 passages of human prose.

The task-fit checks are simple rules (a stance word near the start of an opinion piece, first-person pronouns in a reflection, a greeting in an email, word counts) and they miss things a reader would catch. The training data was filtered on the same checks reported here, so the task-fit numbers flatter the model somewhat.

The neuron edit flips the sign of four MLP neurons (layer 32 neuron 4852, layer 33 neuron 348, layer 28 neuron 1121, layer 32 neuron 5418), which is the same as negating four columns of their layers' down-projection weights. The GGUF build was quantised from a Q8_0 conversion of the merged weights.

The readers' chat and live cursors connect readers directly over WebRTC (Trystero with public Nostr relays). Peers can see each other's IP addresses, nothing is saved, and "hide" disconnects you.