Official Pytorch implementation of MuCo: Multi-turn Contrastive Learning for Multimodal Embedding Model (CVPR 2026)

🚀 First survey on Attention Sink in Transformers — 180+ papers on utilization, interpretation, and mitigation.

Use Codex from Claude Code to review code or delegate tasks.

[CVPR 2026] Code and Datasets for Paper: Learning Transferable Temporal Primitives for Video Reasoning via Synthetic Videos

Codebase for the work “Why Does Self-Distillation (Sometimes) Degrade the Reasoning Capability of LLMs?”

Track Deep-Thinking Tokens in Transformer Models

AI agents running research on single-GPU nanochat training automatically

A Curated List of Awesome Works in World Modeling, Aiming to Serve as a One-stop Resource for Researchers, Practitioners, and Enthusiasts Interested in World Modeling.

[ICLR 2026] Official implementation of the paper "Map the Flow: Revealing Hidden Pathways of Information in VideoLLMs"

Code for the experiments and websites of the paper "Same Task, Different Circuits"

official implementation of CVPR 2026 paper: Test-Time Alignment of Text-to-Image Diffusion Models via Null-Text Embedding Optimisation (Null-TTA)

Official Code for Cross-Class Feature Augmentation for Class Incremental Learning (AAAI2024)

Paper list of Video LLM hallucination. Welcome to Star and Contribute!

The official implementation for the paper "What does RL improve for Visual Reasoning? A Frankenstein-Style Analysis"

(NeurIPS 2025 🔥) Official implementation for "Efficient Multi-modal Large Language Models via Progressive Consistency Distillation"

Official repo for "PAPO: Perception-Aware Policy Optimization for Multimodal Reasoning"

Provide with pre-build flash-attention 2 and 3 package wheels on Linux and Windows using GitHub Actions

Diagnose and Fix CUDA / GPU environments compatibility issues locally, in Docker, and CI/CD. CLI + MCP server available.

[TMLR 2025] Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models

[NeurIPS 2025] Official PyTorch implementation of "Token Bottleneck: One Token to Remember Dynamics"

TinyLLaVA-Video-R1: Towards Smaller LMMs for Video Reasoning

The Github repo for our survey paper: "Locate, Steer, and Improve: A Practical Survey of Actionable Mechanistic Interpretability in Large Language Models"

[CVPR2026] VideoAuto-R1: Video Auto Reasoning via Thinking Once, Answering Twice

Code for EMNLP25 paper "Video-RTS: Rethinking Reinforcement Learning and Test-Time Scaling for Efficient and Enhanced Video Reasoning"

Video-R1: Reinforcing Video Reasoning in MLLMs [🔥the first paper to explore R1 for video]