Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning
Paper
•
2506.01939
•
Published
•
151
An Yang
yangapku
AI & ML interests
NLP and Deep Learning
Recent Activity
authored
a paper
7 days ago
Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective
Reinforcement Learning for LLM Reasoning
authored
a paper
7 days ago
Rationales Are Not Silver Bullets: Measuring the Impact of Rationales on
Model Performance and Reliability
authored
a paper
7 days ago
Qwen3 Embedding: Advancing Text Embedding and Reranking Through
Foundation Models
Organizations
yangapku's activity
authored
3
papers
7 days ago
authored
7
papers
28 days ago
Rethinking Data Selection at Scale: Random Selection is Almost All You Need
Paper
•
2410.09335
•
Published
•
17
Language Models can Self-Lengthen to Generate Long Texts
Paper
•
2410.23933
•
Published
•
18
Qwen2.5 Technical Report
Paper
•
2412.15115
•
Published
•
368
CodeElo: Benchmarking Competition-level Code Generation of LLMs with Human-comparable Elo Ratings
Paper
•
2501.01257
•
Published
•
53
Qwen2.5-1M Technical Report
Paper
•
2501.15383
•
Published
•
71
WorldPM: Scaling Human Preference Modeling
Paper
•
2505.10527
•
Published
•
33
Qwen3 Technical Report
Paper
•
2505.09388
•
Published
•
193
upvoted
2
papers
28 days ago
updated
a
model
about 1 month ago
updated
a
collection
about 1 month ago