Open Source Deep Research Alternative to Reason and Search on Private Data. Written in Python.

Simple extension on vLLM to help you speed up reasoning model without training.

MedReason: Eliciting Factual Medical Reasoning Steps in LLMs via Knowledge Graphs

Lightweight replication study of DeepSeek-R1-Zero. Interesting findings include "No Aha Moment", "Longer CoT ≠ Accuracy", and "Language Mixing in Instruct Models".

Pure RL to post-train base models for social reasoning capabilities. Lightweight replication of DeepSeek-R1-Zero with Social IQa dataset.

xVerify: Efficient Answer Verifier for Reasoning Model Evaluations

Agentic Deep Graph Reasoning Implementation

State Sandbox is an experimental game for socioeconomic simulation. It uses Large Language Models (o3-mini) to simulate the world and complex policy impacts.

Sudoku4LLM is a Sudoku dataset generator for training and evaluating reasoning in Large Language Models (LLMs). It offers customizable puzzles, difficulty levels, and 11 serialization formats to support structured data reasoning and Chain of Thought (CoT) experiments.

Replication Materials for, "Implicit Bias-Like Patterns in Reasoning Models"

LLM reasoning method: combining reflexive cueing with real-time web search and multi-stage analysis for more accurate and explainable answers.

AI Lawyer is an intelligent reasoning legal assistant powered by DeepSeek , Ollama RAG and LangChain, designed to streamline legal research and document analysis. By leveraging retrieval-augmented generation (RAG), it provides precise legal insights, and contract summarization. With an intuitive Streamlit-based UI, analyze legal documents.

Experiments on test-time scaling approaches for reasoning LM's to enforce better <think> or <wait> capabilities.

Model Mondays is a weekly livestreamed series on Microsoft Reactor that helps you make informed model choice decisions with timely updates and model deep-dives. Watch live for the content. Join Discord for the discussions.

xVerify: Efficient Answer Verifier for Large Language Model Evaluations