Skill-of-Mind-Infused LLM
Young-Jun Lee PRO
passing2961
AI & ML interests
Social Dialogue System, Multi-Modal Dialogue
Recent Activity
upvoted
a
paper
3 days ago
Cosmos World Foundation Model Platform for Physical AI
upvoted
a
paper
4 days ago
PRMBench: A Fine-grained and Challenging Benchmark for Process-Level
Reward Models
upvoted
a
paper
4 days ago
Test-time Computing: from System-1 Thinking to System-2 Thinking
Organizations
None yet
models
8
passing2961/Meteor-Image
Updated
passing2961/Thanos-8B
Updated
•
5
•
2
passing2961/Thanos-3B
Updated
•
73
•
1
passing2961/Thanos-1B
Updated
•
7
passing2961/Ultron-11B
Updated
•
8
•
1
passing2961/Ultron-Summarizer-8B
Updated
•
7
•
1
passing2961/Ultron-Summarizer-3B
Updated
•
5
passing2961/Ultron-Summarizer-1B
Updated
•
6
datasets
9
passing2961/photochat_plus
Viewer
•
Updated
•
968
•
28
•
3
passing2961/multifaceted-skill-of-mind
Viewer
•
Updated
•
100k
•
40
•
2
passing2961/stark-dialogue
Viewer
•
Updated
•
86.9k
•
56
•
1
passing2961/stark-image
Viewer
•
Updated
•
1.72M
•
119
•
3
passing2961/stark-image-url
Viewer
•
Updated
•
899k
•
56
•
1
passing2961/stark-summary
Viewer
•
Updated
•
53.3k
•
61
•
1
passing2961/stark-face-image
Viewer
•
Updated
•
93.6k
•
83
•
2
passing2961/KorEmpatheticDialogues
Viewer
•
Updated
•
24.8k
•
38
•
1
passing2961/dialogcc
Viewer
•
Updated
•
83.4k
•
36
•
8