LatentGym: A Testbed For Cross-Task Experiential Learning With Controllable Latent Structure
cs.LG, cs.AI
Submitted: 2026-06-13
Updated: 2026-11-04
Terminology
Sources
- LMRL Gym: Benchmarks for Multi-Turn Reinforcement Learning with Language Models
- A Tutorial on Meta-Reinforcement Learning
- OpenAI Gym
- SkyRL-Agent: Efficient RL Training for Multi-turn LLM Agent
- DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
- RL$^2$: Fast Reinforcement Learning via Slow Reinforcement Learning
- TextArena
- When Continual Learning Moves to Memory: A Study of Experience Reuse in LLM Agents
- Scaling In-Context Online Learning Capability of LLMs via Cross-Episode Meta-RL
- GEM: A Gym for Agentic LLMs
- The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery
- Meta-RL Induces Exploration in Language Agents
- EVOLvE: Evaluating and Optimizing LLMs For In-Context Exploration
- Open-Ended Learning Leads to Generally Capable Agents
- GPT-4o System Card
- BALROG: Benchmarking Agentic LLM and VLM Reasoning On Games
- SWE-Bench-CL: Continual Learning for Coding Agents
- Proximal Policy Optimization Algorithms
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- Learning to reinforcement learn
Related papers
- A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis