29% of Atlas Score
This repository hosts and visualizes results from the Disposition benchmark, providing tools to validate submissions and generate comparative charts of language models across seven dispositional traits.
50 yr old disabled guy, Used to make games, then I made crypto, now I make AI. AKA @jerrimus & @pennykoin
AI-Atlas score
| Machine Learning | 57 |
|---|---|
| Computer Vision | 0 |
| Data Science | 39 |
| NLP & LLMs | 64 |
| AI Agents | 66 |
| MLOps | 29 |
Each node is one domain, scored only from evidence in these public repositories. Select one to see why.
Deep, repeated, well-engineered work in this domain.
51 public repositories → 18 with AI work, evaluated. The other 33 (repositories without AI code, forks and empty repositories) are not scored and have no effect on the result. How repositories are selected
Ordered by how much each project adds to the Atlas score, not by its own score. A project with evidence in several domains can rank above one with a slightly higher project score. Choose a domain to focus on the projects with meaningful evidence there.
29% of Atlas Score
This repository hosts and visualizes results from the Disposition benchmark, providing tools to validate submissions and generate comparative charts of language models across seven dispositional traits.
18% of Atlas Score
Enclosure is a behavioral scenario harness for evaluating AI models' tool use and interactive behavior in controlled, reproducible environments using simulated tools and scripted agents.
16% of Atlas Score
Domicile is a continuous agent system with a persistent memory and tool usage capabilities, designed to run locally. It includes a training component using GRPO to improve the agent's performance on custom-generated scenarios.
13% of Atlas Score
GemDesk is a desktop application that leverages Google's Gemini 3 Flash model to enable multi-file analysis, including document conversion, question answering, chart generation, and report generation via a graphical user interface.
10% of Atlas Score
Disposition is a benchmarking tool for evaluating tool-using AI models across seven behavioral traits (curiosity, creativity, patience, accommodation, hubris, sycophancy, and instruction following). It relies on the Enclosure harness to run scenarios and provides grading logic to assess model performance.
7% of Atlas Score
ERNIE Desktop is a local AI desktop application that integrates llama.cpp for LLM inference, a FastAPI server for web search and system telemetry, and a web-based user interface. It is designed to run on Linux and single-board computers, providing an offline-capable assistant with real-time web search augmentation.