Daily Papers of 2024-04-16

1. Learn Your Reference Model for Real Good Alignment 75 upvotes, #1 of 2024-04-16
2. Megalodon: Efficient LLM Pretraining and Inference with Unlimited Context Length 52 upvotes, #2 of 2024-04-16
3. TransformerFAM: Feedback attention is working memory 37 upvotes, #3 of 2024-04-16
4. Video2Game: Real-time, Interactive, Realistic and Browser-Compatible Environment from a Single Video 26 upvotes, #4 of 2024-04-16
5. Compression Represents Intelligence Linearly 26 upvotes, #4 of 2024-04-16
6. Ctrl-Adapter: An Efficient and Versatile Framework for Adapting Diverse Controls to Any Diffusion Model 16 upvotes, #6 of 2024-04-16
7. HQ-Edit: A High-Quality Dataset for Instruction-based Image Editing 11 upvotes, #7 of 2024-04-16
8. Tango 2: Aligning Diffusion-based Text-to-Audio Generations through Direct Preference Optimization 10 upvotes, #8 of 2024-04-16
9. TextHawk: Exploring Efficient Fine-Grained Perception of Multimodal Large Language Models 9 upvotes, #9 of 2024-04-16
10. On Speculative Decoding for Multimodal Large Language Models 8 upvotes, #10 of 2024-04-16
11. CompGS: Efficient 3D Scene Representation via Compressed Gaussian Splatting 6 upvotes, #11 of 2024-04-16
12. Taming Latent Diffusion Model for Neural Radiance Field Inpainting 6 upvotes, #11 of 2024-04-16

Data: hysts-bot-data/daily-papers-stats and the Daily Papers API. Open data: tardellirs/paper-pulse-data. Sister project: Model Pulse, the download history of every model on the Hub.