Stars
Investor’s Last Exam: Can AI Agents Do Equity Investment Without Seeing the Future?
Ongoing research training transformer models at scale
RSIH — versionable, shareable agent harness: Pi coding agent + Genome config layer
Jev-like family of decision models built on top of Qwen3.5/3.8 you can train and run on your own
DeepSeek Harness: Everything is a Plugin.
[ACL 2025] The code repository for "Mitigating Visual Forgetting via Take-along Visual Conditioning for Multi-modal Long CoT Reasoning" in PyTorch.
ACL'2025: SoftCoT: Soft Chain-of-Thought for Efficient Reasoning with LLMs. and preprint: SoftCoT++: Test-Time Scaling with Soft Chain-of-Thought Reasoning
[CVPR 2026] Machine Mental Imagery: Empower Multimodal Reasoning with Latent Visual Tokens
Claude Code skill: translate English PDFs to Chinese via pdf2zh, preserving layout/formulas/figures.
Awesome-Long2short-on-LRMs is a collection of state-of-the-art, novel, exciting long2short methods on large reasoning models. It contains papers, codes, datasets, evaluations, and analyses.
GitHub Pages site for the Next-Generation AI Systems paper
[TMLR] Rethinking On-Policy Distillation of Large Language Models: Phenomenology, Mechanism, and Recipe
Official codebase for the paper Latent Visual Reasoning
Fully open reproduction of DeepSeek-R1
[ECCV 2026] Demystifying Video Reasoning
[ACL 2026] Render-of-Thought: Rendering Textual Chain-of-Thought as Images for Visual Latent Reasoning
[ICML 2026 & EMNLP 2026] Multimodal deep-research MLLM and benchmark. The first long-horizon multimodal deep-research MLLM, extending the number of reasoning turns to dozens and the number of searc…
[AAAI'2026] Let's Think with Images Efficiently! An Interleaved-Modal Chain-of-Thought Reasoning Framework with Dynamic and Precise Visual Thoughts
[ICML 2026] Official implementation of "Open-o3 Video: Grounded Video Reasoning with Explicit Spatio-Temporal Evidence"
一个用于在 macOS 上平滑你的鼠标滚动效果或单独设置滚动方向的小工具, 让你的滚轮爽如触控板 | A lightweight tool used to smooth scrolling and set scroll direction independently for your mouse on macOS
Open-source evaluation toolkit of large multi-modality models (LMMs), support 220+ LMMs, 80+ benchmarks
OpenCompass is an LLM evaluation platform, supporting a wide range of models from OpenAI, Anthropic, Gemini, Qwen, GLM, DeepSeek, etc, across 100+ datasets covering knowledge, reasoning, coding, sc…
🔥🔥🔥 Latest Papers, Codes and Datasets on Video-LMM Post-Training