A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Jy-oti Aneja, Michael Harrison, Neel Joshi, et al.
Phi-4-reasoning-vision-15B is a compact open-weight multimodal reasoning model demonstrating that careful architecture and data curation enable competitive performance with significantly less compute.
Unknown
This survey systematically reviews the data foundations of long-context language models, covering data sources, curation strategies, and their impact on model performance.
Hongyi Jin, Wenhan Yang, Meysam Ghaffari, et al.
Shows that diverse and challenging reasoning examples for SFT can be identified using only initial reasoning tokens, enabling efficient data curation without strong reasoning models.
Lin Long, Rui Wang, Rui Xiao, et al.
This paper organizes LLM-driven synthetic data generation studies into a unified workflow, highlighting gaps and future research directions.
Unknown
SigLIP 2 improves multilingual vision-language encoders with captioning pretraining, self-supervised losses, and online data curation, offering native aspect ratio preservation.
Unknown
Microsoft Research's Phi-2 demonstrates that small language models can achieve remarkable performance through careful data curation and training, challenging the dominance of large models.