Search papers, labs, and topics across Lattice.
Peking University
4
0
7
ENVS not only reduces training costs but also enhances robustness in GUI tasks, achieving a remarkable pass rate despite real-world interruptions.
Gemini Embedding 2's unified multimodal embeddings beat specialized models across diverse tasks and even generalize zero-shot to niche fields like astronomy and culinary arts.
By explicitly modeling and calibrating a model's intrinsic uncertainty, EGPO unlocks significant gains in reasoning performance for RL-trained language models.
LLMs can now explore knowledge graphs on their own, discovering better reasoning paths and outperforming even closed-source models on question answering.
