TxAgent: An AI Agent for Therapeutic Reasoning Across a Universe of Tools

Project Page For "Seg-Zero: Reasoning-Chain Guided Segmentation via Cognitive Reinforcement"

ToolUniverse is a collection of biomedical tools designed for AI agents

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

Official Implementation of "Reasoning Language Models: A Blueprint"

A Comprehensive Survey on Evaluating Reasoning Capabilities in Multimodal Large Language Models.

This is the repo of developing reasoning models in the specific domain of financial, aim to enhance models capabilities in handling financial reasoning tasks.

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

🔥🔥🔥Breaking long thought processes of o1-like LLMs, such as DeepSeek-R1, QwQ

OpenVLThinker: An Early Exploration to Vision-Language Reasoning via Iterative Self-Improvement

Reasoning-from-Zero using gemma.JAX.nnx on TPUs

Structured test tasks and model tuning scripts for multiple subjects from ZNO - the Ukrainian External Independent Evaluation (ЗНО)

📖Curated list about reasoning abilitiy of MLLM, including OpenAI o1, OpenAI o3-mini, and Slow-Thinking.

This repository contains the implementation of our research on optimizing Retrieval-Augmented Generation (RAG) systems for technical domains. Our work addresses the unique challenges of precise information extraction from complex, domain-specific documents by introducing token-aware evaluation metrics and synthetic data generation pipeline.

Official code for "Divide and Translate: Compositional First-Order Logic Translation and Verification for Complex Logical Reasoning", ICLR 2025.

A curated list of awesome open-source and open-weight language models or methods focused on reasoning capabilities.

AI Model Generator