Where to find Grokking in LLM Pretraining? Monitor Memorization-to-Generalization without Test
Paper
•
2506.21551
•
Published
•
25
•
2
Tianyi Zhou
zhoutianyi
AI & ML interests
ML, NLP, RL, Multi-modality
Recent Activity
commented on
a paper
1 day ago
Where to find Grokking in LLM Pretraining? Monitor
Memorization-to-Generalization without Test
authored
a paper
1 day ago
Don't Think Longer, Think Wisely: Optimizing Thinking Dynamics for Large
Reasoning Models
Organizations
commented
a
paper
1 day ago
commented
3
papers
2 days ago
FaSTA$^*$: Fast-Slow Toolpath Agent with Subroutine Mining for Efficient Multi-turn Image Editing
Paper
•
2506.20911
•
Published
•
37
•
2
FaSTA$^*$: Fast-Slow Toolpath Agent with Subroutine Mining for Efficient Multi-turn Image Editing
Paper
•
2506.20911
•
Published
•
37
•
2
Where to find Grokking in LLM Pretraining? Monitor Memorization-to-Generalization without Test
Paper
•
2506.21551
•
Published
•
25
•
2
commented
a
paper
11 days ago
commented
8
papers
2 months ago
WALL-E 2.0: World Alignment by NeuroSymbolic Learning improves World Model-based LLM Agents
Paper
•
2504.15785
•
Published
•
19
•
4
WALL-E 2.0: World Alignment by NeuroSymbolic Learning improves World Model-based LLM Agents
Paper
•
2504.15785
•
Published
•
19
•
4
Exploring Expert Failures Improves LLM Agent Tuning
Paper
•
2504.13145
•
Published
•
12
•
4
Exploring Expert Failures Improves LLM Agent Tuning
Paper
•
2504.13145
•
Published
•
12
•
4
ColorBench: Can VLMs See and Understand the Colorful World? A Comprehensive Benchmark for Color Perception, Reasoning, and Robustness
Paper
•
2504.10514
•
Published
•
47
•
4
ColorBench: Can VLMs See and Understand the Colorful World? A Comprehensive Benchmark for Color Perception, Reasoning, and Robustness
Paper
•
2504.10514
•
Published
•
47
•
4
ColorBench: Can VLMs See and Understand the Colorful World? A Comprehensive Benchmark for Color Perception, Reasoning, and Robustness
Paper
•
2504.10514
•
Published
•
47
•
4
How Instruction and Reasoning Data shape Post-Training: Data Quality through the Lens of Layer-wise Gradients
Paper
•
2504.10766
•
Published
•
40
•
2
commented
4
papers
3 months ago
C3PO: Critical-Layer, Core-Expert, Collaborative Pathway Optimization for Test-Time Expert Re-Mixing
Paper
•
2504.07964
•
Published
•
61
•
3
C3PO: Critical-Layer, Core-Expert, Collaborative Pathway Optimization for Test-Time Expert Re-Mixing
Paper
•
2504.07964
•
Published
•
61
•
3
Missing Premise exacerbates Overthinking: Are Reasoning Models losing Critical Thinking Skill?
Paper
•
2504.06514
•
Published
•
39
•
3
Missing Premise exacerbates Overthinking: Are Reasoning Models losing Critical Thinking Skill?
Paper
•
2504.06514
•
Published
•
39
•
3
commented
3
papers
4 months ago
CoSTA$\ast$: Cost-Sensitive Toolpath Agent for Multi-turn Image Editing
Paper
•
2503.10613
•
Published
•
80
•
10
CoSTA$\ast$: Cost-Sensitive Toolpath Agent for Multi-turn Image Editing
Paper
•
2503.10613
•
Published
•
80
•
10
CoSTA$\ast$: Cost-Sensitive Toolpath Agent for Multi-turn Image Editing
Paper
•
2503.10613
•
Published
•
80
•
10