#qwen articles | Hashnode

Self-hosting Qwen 3.8 27B

Self-hosted Qwen 3.8 27B on a GPU pod with 2x Nvidia RTX 3090 GPUs and vLLM.

I possessed an LLM by editing its hidden layers, then exorcised it

Activation steering on Qwen 2.5 7B, an exorcist that doesn't need to know the dose, and the day my model graded itself too kindly

5 min readbuild-mcp-with-ai.hashnode.dev

Shivam KumarNov 1

The Qwen Deal Is a Playbook: Local Models Are Becoming the Price of Global Distribution

Alibaba announced on October 31, 2026 that Qwen will power Apple Intelligence in China, with a developer beta on November 4 and the official iOS 27.3 release in December. The headlines will call it an

6 min readshivaysinghrajput.hashnode.dev

Avneesh JadhavOct 28

AI on Edge

Deployed d3-nano and Qwen 3.5 on my own infrastructure using llama.cpp. Run System 1 & 2 models on low resources for unlimited, low-cost inference.

1 min readavneeshjadhav.hashnode.dev

Martin BowlingOct 27

The Celebrity Neuron Is Not a Neuron

I tried to make Qwen3.5-4B identify as Einstein, Curie, Newton, Taylor Swift, and Keanu Reeves. The scientists worked. The entertainers didn’t.

5 min readmartinbowling.com

Martin BowlingOct 26

I Audited the Viral Qwen Neuron Zoo

I tested the viral one-neuron personas on Qwen3.5-4B. 16 of 22 survived a one-digit fix, one was impossible, and three neurons become the Golden Gate Bridge.

10 min readmartinbowling.com

Martin BowlingOct 26

Three-Neuron Bangers

Five combos that fuse into something none of the neurons does alone — plus two cursed loops and one random control that does nothing.

5 min readmartinbowling.com

HexGrid.cloudOct 16

Qwen Image 2.1 Edit: 8 Real-World Image Editing Stress Tests in ComfyUI

Explore Qwen Image 2.1 editing with 9 real-world prompt tests in ComfyUI. See image outputs, editing quality, GPU performance, and inference costs.

5 min readblog.hexgrid.cloud

2

RONI DASOct 15

There is no 4B Qwen 3.8. The smallest open model is 27B, and here is what it took to run it on one rented GPU.

Qwen 3.8 has no 2B, 3B or 4B model. I pulled qwen3.8:27b into Ollama on a rented NVIDIA L4: 17.7 GB on disk, 18,004 MiB of GPU memory in use, 18 to 35 tokens a second once loaded.

12 min readsystemdesignacademy.hashnode.dev

1

Lưu Tuấn AnhOct 6

CLOSEDQUORUM: When 4 AI models turn into C2 servers for Malware

Overview Imagine a system intrusion where the mastermind is not sitting in front of the screen, is not typing any C2 control commands, and is even... sleeping. Instead, the four leading commercial art

11 min readblog.fiscybersec.com

Search Hashnode

Trending tags this week