AI Developer Tool · Python

sub-zero

Rick@JuiceB0xC0de · Atlas 73.3 · #2= builder

What it is

Sub-Zero is an end-to-end interpretability and steering system for LLMs that identifies and attenuates corporate/refusal directions in model weights through SVD-based probing and gradient-masked fine-tuning, accompanied by a pipeline for building a detailed Brain Atlas.

Scored from the code at its last scan, 22 Oct 2026. How merit is scored

Where it counts

NLP & LLMs95%
Machine Learning90%
Data Science40%
MLOps20%
Computer Vision10%
AI Agents10%

How much of the project is about each field, as scoring counts it. It ranks in the field it is most about.

Around it on the NLP & LLMs board

Whole board

RankProjectBuilderFieldMerit

11EuchreGameDataset / Data Pipeline · PythonThe EuchreGame repository provides a deterministic Python Euchre engine, a browser-based sandbox for gameplay and AI interaction via Pyodide, and a data factory pipeline for generating synthetic trajectories to distill a language model into a Euchre-playing agent.63.4/100AdvancedJarad PeckenpaughNLP & LLMs
12llmastsearchAI Developer Tool · RustThe llmastsearch project is a Rust-based command-line tool and library for performing syntax-aware code queries using tree-sitter parsers, providing structured JSON output optimized for LLM consumption. It supports seven programming languages and includes features such as…62.0/100AdvancedLuiz Cesar SpiesNLP & LLMs
13sub-zeroAI Developer Tool · Python61.7/100AdvancedRickNLP & LLMs
14deep-chaos-schedulerArchitecture Modification / Research · PythonThe repository introduces a sticky-topology chaos scheduler (DeepChaos) for transformer fine-tuning that removes dead and identity layers during training to improve speed and reduce VRAM usage, paired with a diagnostic suite (BoL scans) for model analysis.61.0/100AdvancedRickNLP & LLMs
15gemma-servant-voice-suppression-tuningTraining Experiment · PythonThe repository conducts experiments to steer the voice of Gemma language models toward the Bella persona while evaluating impacts on safety capabilities, using direction fitting, steering grids, and LLM-based evaluation.60.8/100AdvancedRickNLP & LLMs

What it is

Where it counts

Around it on the NLP & LLMs board

More from Rick