Reinforcement learning, explained with a minimum of math and jargon

To create reliable agents, AI companies had to go beyond predicting the next token.

Q-learning is not yet scalable

Using Reinforcement Learning to Train Large Language Models to Explain Human Decisions

A central goal of cognitive modeling is to develop models that not only predict human behavior but also provide insight into the underlying cognitive mechanisms. While neural network models trained on large-scale behavioral data often achieve strong predictive performance, they typically fall short in offering interpretable explanations of the cognitive processes they capture. In this work, we explore the potential of pretrained large language models (LLMs) to serve as dual-purpose cognitive models--capable of both accurate prediction and interpretable explanation in natural language. Specifically, we employ reinforcement learning with outcome-based rewards to guide LLMs toward generating explicit reasoning traces for explaining human risky choices. Our findings demonstrate that this approach produces high-quality explanations alongside strong quantitative predictions of human decisions.

Outcome-based Reinforcement Learning to Predict the Future

Reinforcement learning with verifiable rewards (RLVR) has boosted math and coding in large language models, yet there has been little effort to extend RLVR into messier, real-world domains like forecasting. One sticking point is that outcome-based reinforcement learning for forecasting must learn from binary, delayed, and noisy rewards, a regime where standard fine-tuning is brittle. We show that outcome-only online RL on a 14B model can match frontier-scale accuracy and surpass it in calibration and hypothetical prediction market betting by adapting two leading algorithms, Group-Relative Policy Optimisation (GRPO) and ReMax, to the forecasting setting. Our adaptations remove per-question variance scaling in GRPO, apply baseline-subtracted advantages in ReMax, hydrate training with 100k temporally consistent synthetic questions, and introduce lightweight guard-rails that penalise gibberish, non-English responses and missing rationales, enabling a single stable pass over 110k events. Scaling ReMax to 110k questions and ensembling seven predictions yields a 14B model that matches frontier baseline o1 on accuracy on our holdout set (Brier = 0.193, p = 0.23) while beating it in calibration (ECE = 0.042, p < 0.001). A simple trading rule turns this calibration edge into \$127 of hypothetical profit versus \$92 for o1 (p = 0.037). This demonstrates that refined RLVR methods can convert small-scale LLMs into potentially economically valuable forecasting tools, with implications for scaling this to larger models.

INTELLECT-2 Release: The First 32B Parameter Model Trained Through Globally Distributed Reinforcement Learning

We're excited to release INTELLECT-2, the first 32B parameter model trained via globally distributed reinforcement learning. Unlike traditional centralized training efforts, INTELLECT-2 trains a reasoning language model using fully asynchronous RL across a dynamic, heterogeneous swarm of permissionless compute contributors.

Absolute Zero Reasoner

GitHub - ivanbelenky/RL: R.L. methods and techniques.

R.L. methods and techniques. . Contribute to ivanbelenky/RL development by creating an account on GitHub.

GitHub - OpenPipe/ART: OpenPipe ART (Agent Reinforcement Trainer): train LLM agents

OpenPipe ART (Agent Reinforcement Trainer): train LLM agents - OpenPipe/ART

GitHub - deepseek-ai/DeepSeek-Prover-V2

Contribute to deepseek-ai/DeepSeek-Prover-V2 development by creating an account on GitHub.

GitHub - policy-gradient/GRPO-Zero

Contribute to policy-gradient/GRPO-Zero development by creating an account on GitHub.