SofT-GRPO: Surpassing Discrete-Token LLM Reinforcement Learning via Gumbel-Reparameterized Soft-Thinking Policy Optimization
Paper
•
2511.06411
•
Published
•
17
Guyu
kuangrepi
·
AI & ML interests
None yet
Recent Activity
upvoted
a
paper
9 months ago
TextCrafter: Accurately Rendering Multiple Texts in Complex Visual
Scenes
Organizations
upvoted
a
paper
about 2 months ago