Stars
Character Creation Through Strategic Ablation — Permanently ablating personality steering vectors into LLM weights using contrastive activation extraction and benchmark-driven optimization
Steering-first neuron explorer for Qwen3.5-4B: push every MLP neuron, keep every rollout, browse the results
Reverse engineer anything with agents, from app behavior down to native binaries.
OBLITERATE THE CHAINS THAT BIND YOU
My OpenWRT fork for the GL.INET X3000 (Spitz AX)
Quectel RM520N-GL is 5G IoT module specially optimized for IoT/eMBB applications.
Configuration files for AI coding tools, managed with symlinks.
📄 Configuration files that enhance Cursor AI editor experience with custom rules and behaviors
Interactive educational web for karpathy/microgpt.py
High-performance async runtime for Nim with structured concurrency, lock-free channels, and work-stealing scheduler
A tool for building the Linux kernel or other software with many different compilers in Docker/Podman containers
A concurrent, spsc ring-buffer with sized reservations
A SPSC, lockless, no_std, thread safe, queue, based on BipBuffers
A fast multi-producer, multi-consumer lock-free concurrent queue for C++11
"Lock-Free Ring Buffer" (LFRB) is a minimal, customizable implementation of a ring buffer (a.k.a. circular buffer) in C, specifically suitable for embedded systems.
Amiga Minimig ported to the Tang Nano 20k FPGA
Suno AI's Bark model in C/C++ for fast text-to-speech generation
Minimal, clean code for the Byte Pair Encoding (BPE) algorithm commonly used in LLM tokenization.
Official implementation of Cold-Diffusion for different transformations in pytorch.
Nim macros to aid in object construction including event programming, and constructors.
Nim macros that might help writing simpler and safer code.
Print is a set of pretty print macros, useful for print-debugging.