{
  "id": 6647,
  "url": "https://arxiv.org/abs/2604.08579v1",
  "title": "On the Spectral Geometry of Cross-Modal Representations: A Functional Map Diagnostic for Multimodal Alignment",
  "summary": "We study cross-modal alignment between independently pretrained vision (DINOv2) and language (all-MiniLM-L6-v2) encoders using the functional map framework from computational geometry, which represents correspondence between representation manifolds as a compact linear operator between graph Laplacian eigenbases. While the framework underperforms Procrustes alignment and relative representations for cross-modal retrieval across all supervision budgets, it reveals a structural property of multimo",
  "authors": "Krisanu Sarkar",
  "category": "research",
  "topics": "safety-alignment,healthcare",
  "orgs": null,
  "regions": null,
  "published_at": "2026-03-28T04:20:54.000Z",
  "fetched_at": "2026-07-14T16:32:37.311Z",
  "source_slug": "arxiv-ethics",
  "source_name": "arXiv",
  "source_homepage": "https://arxiv.org",
  "ethics_ai_record_url": "https://ethics.ai/record/6647",
  "original_url": "https://arxiv.org/abs/2604.08579v1",
  "evidence_status": "source-only",
  "attribution": "via ethics.ai"
}