A collective of passionate students pushing the boundaries of artificial intelligence research.
Exploring new ideas at the intersection of language, reasoning, and intelligence.
Investigating how transformer architectures encode and retrieve semantic information across layers.
Developing novel sparse attention mechanisms that maintain reasoning quality while reducing computational cost.
Building bridges between visual and linguistic understanding for more holistic AI systems.
Tools and applications that make AI more accessible and useful.
An open platform for running, sharing, and reproducing AI experiments — built by researchers, for researchers.
A next-generation search tool powered by deep semantic understanding, going beyond keyword matching.
We believe in building in the open. Here are some of our contributions.
A comprehensive benchmark suite for evaluating language model capabilities across reasoning, retrieval, and generation tasks.
A lightweight library for building semantic search pipelines with pluggable embedding models and vector stores.
Interactive visualization tools for exploring neural network activations, attention patterns, and learned representations.
Passionate students from diverse backgrounds united by a love for AI.