REVE: Efficient Hallucination Correction for Large Audio-Language Models via Reused Encoder States
This work proposes Reused Encoder States for Verifying Events (REVE), a lightweight method that uses states already computed by the target model that achieves a reduction comparable to those of CED-Tiny and CED-Base.
Hong-Jin Song, Ji-Shen Kuang, Xin-Yu Yang et al.
· 0 citations