A general framework for strategically scaling evaluation-driven discovery loops, discovering state-of-the-art solutions on 21 open-ended problems.

Reimplementation of TTT-Discover without the Tinker dependency

CaOPD: Calibration-Aware On-Policy Distillation

Lightweight coding agent that runs in your terminal

Bash Line Editor―a line editor written in pure Bash with syntax highlighting, auto suggestions, vim modes, etc. for Bash interactive sessions.

AI-Driven Scientific and Algorithmic Discovery

Framework for evaluating and improving agents

OpenClaw-RL: Train any agent simply by talking

The official codebase for "Experiential Reinforcement Learning" - https://arxiv.org/pdf/2602.13949v1

PyTorch library for Active Fine-Tuning

Aligning Language Models from User Interactions via Self-Distillation

Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞

slime is an LLM post-training framework for RL Scaling.

Continual Learning as a Service

pip install continualcode

Reinforcement Learning via Self-Distillation (SDPO)

A high-throughput and memory-efficient inference and serving engine for LLMs

Official JAX implementation of End-to-End Test-Time Training for Long Context

Hydra is a framework for elegantly configuring complex applications

Storing long contexts in tiny caches with self-study

verl/HybridFlow: A Flexible and Efficient RL Post-Training Framework

Test-Time Curricula for Targeted RL

Specialization after Generalization

The official repository of paper "Pass@k Training for Adaptively Balancing Exploration and Exploitation of Large Reasoning Models''

[ICLR 2026] On the Generalization of SFT: A Reinforcement Learning Perspective with Reward Rectification.

Official repository for "Test-time Offline Reinforcement Learning on Goal-related Experience" (Preprint)

Hierarchical Reasoning Model Official Release

The codebase and some introductions of FineMed.