Search papers, labs, and topics across Lattice.
3
0
5
4
LVLMs are better at spotting their own mistakes than generating correct answers in the first place, and this self-awareness can be exploited to reduce hallucinations.
By co-evolving experts through bidirectional policy distillation, CoPD achieves all-in-one integration of text, image, and video reasoning, outperforming domain-specific experts and suggesting a new training paradigm.
Self-distillation in LLMs can leak information and destabilize training, but combining it with verifiable rewards yields a sweet spot for improved convergence and stability.
