A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Unknown
Investigates automated data selection for instruction-tuning LLMs, finding many methods underperform random selection at scale, while a representation-based variant (RDS+) consistently outperforms others.
Unknown
CoSyn uses LLMs to generate code that renders synthetic images, then uses that code as context to produce high-quality instruction-tuning data for vision-language models.
Unknown
MAmmoTH-VL introduces a cost-effective pipeline to create a 12M multimodal instruction-tuning dataset with chain-of-thought reasoning, achieving state-of-the-art performance on diverse benchmarks.