GRPO reinforcement learning for LLMs on math problems with verifiable reward — trained on Ascend NPU with verl
-
Updated
Aug 30, 2026
#
verifiable-reward
Here are 2 public repositories matching this topic...
A verifiable benchmark + RL environment for agents that design neural-network architectures. Models emit graph edits; a deterministic verifier grades the result. No GPU, no human judge.
benchmark reinforcement-learning deep-learning grok rl-environment neural-architecture evals llm-agents grpo verifiable-reward
-
Updated
Sep 2, 2026 - JavaScript
Add this topic to your repo
To associate your repository with the verifiable-reward topic, visit your repo's landing page and select "manage topics."