Horizon Team ORG

[ICML2026] Reproduce Kimi K1.5/K2 RL algorithm and rollout system

[NeurIPS 2025] Think-RM: Enabling Long-Horizon Reasoning in Generative Reward Models

[NeurIPS 2025] Ask a Strong LLM Judge when Your Reward Model is Uncertain

[COLM2026] Difficulty-Aware Heap Sampling and On-Policy Query Augmentation for LLM Reinforcement Learning

[arXiv 2026] Learning from Rare Success and Rich Feedback via Reflection-Enhanced Self-Distillation

Rethinking Importance Sampling in LLM Policy Optimization: A Cumulative Token Perspective

[ACL2026] AlphaQuanter: An End-to-End Tool-Orchestrated Agentic Reinforcement Learning Framework for Stock Trading.

[ICML 2026] CoMem: Context Management with A Decoupled Long-Context Model

Multi-Step Tool Orchestration with Constrained Data Synthesis and Graduated Rewards

Implementation for paper Improving Sampling Efficiency in RLVR through Adaptive Rollout and Response Reuse

No description.