Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning
Paper
•
2506.01939
•
Published
•
129
Robin Williams PRO
bfuzzy1
AI & ML interests
None yet
Recent Activity
updated
a collection
11 days ago
Nifty
upvoted
a
paper
11 days ago
Shifting AI Efficiency From Model-Centric to Data-Centric Compression
Organizations
None yet
bfuzzy1's activity
upvoted
a
paper
4 days ago
updated
a
collection
11 days ago
upvoted
5
papers
11 days ago
Shifting AI Efficiency From Model-Centric to Data-Centric Compression
Paper
•
2505.19147
•
Published
•
144
Truth Neurons
Paper
•
2505.12182
•
Published
•
7
dKV-Cache: The Cache for Diffusion Language Models
Paper
•
2505.15781
•
Published
•
16
Diffusion vs. Autoregressive Language Models: A Text Embedding Perspective
Paper
•
2505.15045
•
Published
•
53
Position of Uncertainty: A Cross-Linguistic Study of Positional Bias in Large Language Models
Paper
•
2505.16134
•
Published
•
18
updated
a
collection
13 days ago
updated
a
collection
16 days ago
upvoted
a
paper
25 days ago
updated
a
collection
about 1 month ago
upvoted
6
papers
about 1 month ago
X-Fusion: Introducing New Modality to Frozen Large Language Models
Paper
•
2504.20996
•
Published
•
12
Reinforcement Learning for Reasoning in Large Language Models with One Training Example
Paper
•
2504.20571
•
Published
•
94
The Leaderboard Illusion
Paper
•
2504.20879
•
Published
•
70
ReasonIR: Training Retrievers for Reasoning Tasks
Paper
•
2504.20595
•
Published
•
55
Toward Evaluative Thinking: Meta Policy Optimization with Evolving Reward Models
Paper
•
2504.20157
•
Published
•
37
TesserAct: Learning 4D Embodied World Models
Paper
•
2504.20995
•
Published
•
20
updated
a
collection
about 1 month ago
upvoted
a
paper
about 1 month ago