General Reasoning
80 posts
@GenReasoning
A long-horizon reinforcement learning company.
Major FutureSim updates! 🥽 Visualized trajectories: extremely long horizon, 1000s of actions per run and dozens of compactions, with summaries! Goldmine of interesting agent behaviors across Fable 5, GPT 5.5, GLM 5.2 etc! 📈 @OpenReward integration! You can now use FutureSim, incl. on your own data, with our favourite standard! The @GenReasoning was super supportive in making this happen! Some cool observations in 🧵
Excited to be speaking at @aiDotEngineer World’s Fair alongside partner in crime @ChengxiTaylor. It seems we have a lot to talk about: 🤼♂️ Actor-critic is hot again? How are current RL algorithms changing with long-horizon rollouts. 🌍 What does a good long-horizon environment look like? How do they differ from terminal based tasks? 🤖 What does good RL infra look like in 2026? Why is long horizon breaking things? As a bonus, I’ll give a run down of our own RL journey starting from Galactica (2021/2022), early SoTA reasoning efforts at Meta (2023), and more. Super fun, high signal-to-noise guaranteed. Come join us!
You can now train on 350+ RL Environments from OpenReward with TRL with just a few lines of code
