An effective and universal harness for embodied manipulation
Xirui Li PRO
AIcell
AI & ML interests
Multi-Modality
Recent Activity
upvoted a paper about 8 hours ago
4DCodeBench: Benchmarking Agents on Inverse Graphics of Dynamic Scenes upvoted a paper 4 days ago
Decoding Looped Transformers Better for (Almost) Free upvoted a paper 4 days ago
Sharpening Tax in Post-TrainingOrganizations
We us AI to provide war predictions.
The collection for the paper: What does RL improve for Visual Reasoning? A Frankenstein-Style Analysis
-
What does RL improve for Visual Reasoning? A Frankenstein-Style Analysis
Paper • 2602.12395 • Published • 17 -
AIcell/Frankenstein-IN
Image-Text-to-Text • 4B • Updated • 12 -
AIcell/Frankenstein-RL
Image-Text-to-Text • 4B • Updated • 22 -
AIcell/Frankenstein-RL-Frozen-Early
Image-Text-to-Text • 4B • Updated • 16
An effective and universal harness for embodied manipulation
Scalable Environment Generation for Claw-Like Agents
We us AI to provide war predictions.
MoltBook datasets and papers, please check live statistics on https://ai-agent-society.com/.
The collection for the paper: What does RL improve for Visual Reasoning? A Frankenstein-Style Analysis
-
What does RL improve for Visual Reasoning? A Frankenstein-Style Analysis
Paper • 2602.12395 • Published • 17 -
AIcell/Frankenstein-IN
Image-Text-to-Text • 4B • Updated • 12 -
AIcell/Frankenstein-RL
Image-Text-to-Text • 4B • Updated • 22 -
AIcell/Frankenstein-RL-Frozen-Early
Image-Text-to-Text • 4B • Updated • 16
The collection of the paper: "R1-Zero's "Aha Moment" in Visual Reasoning on a 2B Non-SFT Model"