{
  "schema_version": 1,
  "title": "Reproduction: CrossQ: Task-Aligned Cross-Token Conditional Quantization for Late Interaction Retrieval",
  "emoji": "🎯",
  "space_id": "phantomcoder1996/repro-crossq-task-aligned-cross-token-conditional-quantization-for-late-interaction-retrieval",
  "paper": null,
  "tags": [
    "icml2026-repro",
    "paper-jxHF6z0S1L"
  ],
  "updated_at": "2026-07-18T18:35:04+00:00",
  "root": {
    "slug": "index",
    "title": "Reproduction: CrossQ: Task-Aligned Cross-Token Conditional Quantization for Late Interaction Retrieval",
    "file": "pages/index.md",
    "children": [
      {
        "slug": "executive-summary",
        "title": "Executive summary",
        "file": "pages/executive-summary/page.md",
        "children": []
      },
      {
        "slug": "claim-1-late-interaction-retrievers-like-colbert-achieve-high-quality-but-suffer-from-large-multi-vector-indices",
        "title": "Claim 1: Late-interaction retrievers like ColBERT achieve high quality but suffer from large multi-vector indices.",
        "file": "pages/claim-1-late-interaction-retrievers-like-colbert-achieve-high-quality-but-suffer-from-large-multi-vector-indices/page.md",
        "children": []
      },
      {
        "slug": "claim-2-crossq-adaptively-allocates-precision-by-conditioning-token-codes-on-lightweight-document-context-computed-at-indexing-time",
        "title": "Claim 2: CrossQ adaptively allocates precision by conditioning token codes on lightweight document context computed at indexing time.",
        "file": "pages/claim-2-crossq-adaptively-allocates-precision-by-conditioning-token-codes-on-lightweight-document-context-computed-at-indexing-time/page.md",
        "children": []
      },
      {
        "slug": "claim-3-achieves-61x-compression-while-narrowing-gap-with-full-precision-colbert-to-just-2-3-mrr-10-loss",
        "title": "Claim 3: Achieves 61x compression while narrowing gap with full-precision ColBERT to just 2.3% MRR@10 loss.",
        "file": "pages/claim-3-achieves-61x-compression-while-narrowing-gap-with-full-precision-colbert-to-just-2-3-mrr-10-loss/page.md",
        "children": []
      },
      {
        "slug": "conclusion",
        "title": "Conclusion",
        "file": "pages/conclusion/page.md",
        "children": []
      }
    ]
  },
  "agent_view_tokens": 2136,
  "revision": "1784399704185825600"
}