{
  "schema_version": "1.0",
  "id": "archive:https://arxiv.org/abs/2607.11801v1",
  "slug": "2607-11801v1-0m0m7wj",
  "url": "https://feed7.dev/p/2607-11801v1-0m0m7wj",
  "title": "Encoder-Side Neuron Identification and Amplification for Acoustic Perception in Large Audio-Language Models",
  "why_included": "IAAN boosts selected audio-encoder neurons at inference, improving fine-grained speech perception across three models without retraining or labels.",
  "summary": "IAAN scores audio-encoder neurons by contrasting responses to a real waveform and a noise reference, then amplifies a small selected set during inference. It requires **no training or labels**.",
  "practical_implication": "Across ten non-semantic speech attributes, average accuracy rose by **25.7 points on Audio-Flamingo-3**, **21.4 on Qwen2.5-Omni**, and **9.7 on Kimi-Audio**. Builders with encoder access could test targeted activation changes before committing to fine-tuning.",
  "agent_context": "IAAN scores audio-encoder neurons by contrasting responses to a real waveform and a noise reference, then amplifies a small selected set during inference. It requires **no training or labels**.\n\nAcross ten non-semantic speech attributes, average accuracy rose by **25.7 points on Audio-Flamingo-3**, **21.4 on Qwen2.5-Omni**, and **9.7 on Kimi-Audio**. Builders with encoder access could test targeted activation changes before committing to fine-tuning.\n\nThe gains depended on intervening inside the encoder and choosing the right neurons. Post-encoder or language-model interventions offered little benefit or reduced accuracy, and the material does not establish results beyond the tested models and attributes.",
  "source": {
    "name": "arXiv",
    "url": "https://arxiv.org/abs/2607.11801v1",
    "published_at": null
  },
  "source_class": "blog_post",
  "content_type": "Paper",
  "layer": "model",
  "domains": [
    "audio"
  ],
  "topics": [
    "generative-media"
  ],
  "verification": {
    "status": "needs_review",
    "label": "Needs Review",
    "method": "unverified",
    "verified_at": null
  },
  "uncertainty": [
    "The gains depended on intervening inside the encoder and choosing the right neurons. Post-encoder or language-model interventions offered little benefit or reduced accuracy, and the material does not establish results beyond the tested models and attributes."
  ],
  "lifecycle": "Current",
  "published_at": null,
  "modified_at": null,
  "supersedes": [],
  "expires_at": null,
  "formats": {
    "html": "https://feed7.dev/p/2607-11801v1-0m0m7wj",
    "json": "https://feed7.dev/p/2607-11801v1-0m0m7wj.json",
    "markdown": "https://feed7.dev/p/2607-11801v1-0m0m7wj.md"
  }
}