Thinking in Space: How Multimodal LLM See, Remember, and Recall Spaces (Paper Walkthrough)
20:06
Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMs (Paper Walkthrough)
13:53
Rho-1: Not All Tokens Are What You Need -- NeurIPS 2024 Runner-Up (Paper Walkthrough)
14:45
Adam on Local Time: Addressing Nonstationarity in RL with Relative Adam Timesteps (Paper Walkthru)
15:16
ICONS: Influence Consensus for Vision-Language Data Selection (Paper Walkthrough)
44:18
TapeAgents: a Holistic Framework for Agent Development and Optimization
38:49
DeepSeek-V3 Technical Report Walkthrough
22:32
4D Gaussian Splatting: Modeling Dynamic Scenes with Native 4D Primitives (Paper Walkthrough)
15:23