logbook.json

{
  "schema_version": 1,
  "title": "Repro: QuArch - A Benchmark for Evaluating LLM Reasoning in Computer Architecture",
  "emoji": "🎯",
  "space_id": "ParetoOptimal/repro-quarch",
  "paper": {
    "arxiv_id": "2510.22087"
  },
  "tags": [
    "icml2026-repro",
    "paper-yU6X1XZl8t"
  ],
  "updated_at": "2026-07-16T22:16:40+00:00",
  "root": {
    "slug": "index",
    "title": "Repro: QuArch - A Benchmark for Evaluating LLM Reasoning in Computer Architecture",
    "file": "pages/index.md",
    "children": [
      {
        "slug": "paper-approach",
        "title": "Paper & approach",
        "file": "pages/paper-approach/page.md",
        "children": []
      },
      {
        "slug": "official-claim-1-qa-pairs",
        "title": "Official Claim 1 — 2,671 expert-validated QA pairs (composition & coverage)",
        "file": "pages/official-claim-1-qa-pairs/page.md",
        "children": []
      },
      {
        "slug": "official-claim-2-frontier-gaps",
        "title": "Official Claim 2 — Frontier accuracies 34–73%, wide higher-order gaps",
        "file": "pages/official-claim-2-frontier-gaps/page.md",
        "children": []
      },
      {
        "slug": "official-claim-3-finetuning-design",
        "title": "Official Claim 3 — Fine-tuning → 1.99× area-efficiency, 40% more viable",
        "file": "pages/official-claim-3-finetuning-design/page.md",
        "children": []
      },
      {
        "slug": "conclusion",
        "title": "Conclusion",
        "file": "pages/conclusion/page.md",
        "children": []
      }
    ]
  },
  "agent_view_tokens": 7747,
  "revision": "1784240200861629888"
}