{
  "id": 9243510,
  "title": "REVE: Efficient Hallucination Correction for Large Audio-Language Models via Reused Encoder States",
  "url": "https://urgent.news/2026/09/22/reve-efficient-hallucination-correction-for-large-audio-language",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-22T11:32:22.000Z",
  "source": {
    "name": "arXiv cs.AI",
    "slug": "arxiv-cs-ai",
    "url": "https://arxiv.org/abs/2609.26028v1"
  },
  "original_language": "en",
  "account": null,
  "summary": "Large audio-language models may mention acoustic events that are absent from the input. A separate audio event detector can verify these mentions, but doing so requires a second audio encoder and a separate forward pass. We propose Reused Encoder States for Verifying Events (REVE), a lightweight method that uses states already computed by the target model. One readout summarizes class scores…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}