Search papers, labs, and topics across Lattice.
4
7
7
5
95% of emotional support systems prioritize immediate relief over long-term user capability, highlighting a critical gap in current research and practice.
Achieving low regret in contextual procurement auctions reveals a nuanced tradeoff between welfare and incentive errors that could transform auction mechanisms.
Off-policy degree in RLVR updates can drastically change which tokens drive learning, leading to a new adaptive method that outperforms traditional baselines.
Forget fine-tuning: this evolutionary approach lets you adapt LLMs to new tasks with just 200 samples and no gradients, outperforming standard methods by up to 54.8%.
