Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
3
2
7
4
Rare concepts in LLM fine-tuning are significantly hindered by "prior barriers" that require targeted supervision, not just more data.
Meta's new hierarchical indexing method lets you deploy massive recommendation models without sacrificing speed or accuracy, and it turns out the index itself highlights a high-quality subset of data perfect for test-time training.
Forget imbalanced LoRA usage: ReMix leverages reinforcement learning to route effectively among LoRAs, boosting performance in parameter-efficient fine-tuning.