Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning
Paper
•
2506.01939
•
Published
•
177
Wang Yuanqiu
dadaniel
AI & ML interests
None yet
Recent Activity
liked
a model
about 1 month ago
Dream-org/Dream-v0-Instruct-7B
new activity
about 2 months ago
taobao-mnn/InternVL2_5-1B-MNN:Guide on converting original internVL into mnn format
liked
a model
about 2 months ago
taobao-mnn/InternVL2_5-1B-MNN
Organizations
None yet
upvoted
a
paper
2 months ago
upvoted
an
article
3 months ago
Article
DeepSeek-R1 Dissection: Understanding PPO & GRPO Without Any Prior Reinforcement Learning Knowledge
By
•
•
208
upvoted
a
collection
3 months ago