{
  "schema_version": "1.0",
  "id": "archive:https://arxiv.org/abs/2607.05382v1",
  "slug": "2607-05382v1-1xo10v8",
  "url": "https://feed7.dev/p/2607-05382v1-1xo10v8",
  "title": "Search Beyond What Can Be Taught: Evolving the Knowledge Boundary in Agentic Visual Generation",
  "why_included": "SearchGen-Bench shows open image generators score 21–28/100 on long-tail entities, and naive search retrieval only adds noise; a teach-then-search co-training recipe learns when to retrieve versus rely on weights.",
  "summary": "**The gist** **SearchGen-Bench** pairs **20,839 prompts** across twelve failure categories with a pre-executed 1M-item multimodal corpus for offline experiments. Frontier open visual generators score only **21–28 out of 100** — a roughly **40-point collapse** existing benchmarks miss — because they confidently fabricate long-tail entities and post-cutoff events.",
  "practical_implication": "**Why it matters** The retrieval lesson agents keep re-learning applies to image generation: **naive search fails**, injecting noise into prompts the generator already handles. The paper frames a generator-specific **knowledge boundary** — what lives in weights versus external context — and shows a **teach-then-search co-training** loop discovers it, improving monotonically. If you wire image generation into agent pipelines, gate retrieval on that boundary rather than searching every request.",
  "agent_context": "**The gist** **SearchGen-Bench** pairs **20,839 prompts** across twelve failure categories with a pre-executed 1M-item multimodal corpus for offline experiments. Frontier open visual generators score only **21–28 out of 100** — a roughly **40-point collapse** existing benchmarks miss — because they confidently fabricate long-tail entities and post-cutoff events.\n\n**Why it matters** The retrieval lesson agents keep re-learning applies to image generation: **naive search fails**, injecting noise into prompts the generator already handles. The paper frames a generator-specific **knowledge boundary** — what lives in weights versus external context — and shows a **teach-then-search co-training** loop discovers it, improving monotonically. If you wire image generation into agent pipelines, gate retrieval on that boundary rather than searching every request.\n\n**Watch out** Scores cover **open generators** only, and only a **minimal version** of the co-training recipe is demonstrated; whether the learned boundary keeps up as generators and trending entities keep moving is the open question.",
  "source": {
    "name": "arXiv",
    "url": "https://arxiv.org/abs/2607.05382v1",
    "published_at": null
  },
  "source_class": "blog_post",
  "content_type": "Paper",
  "layer": "context",
  "domains": [
    "image",
    "research"
  ],
  "topics": [
    "retrieval",
    "tool-use",
    "generative-media"
  ],
  "verification": {
    "status": "needs_review",
    "label": "Needs Review",
    "method": "unverified",
    "verified_at": null
  },
  "uncertainty": [
    "Scores cover **open generators** only, and only a **minimal version** of the co-training recipe is demonstrated; whether the learned boundary keeps up as generators and trending entities keep moving is the open question."
  ],
  "lifecycle": "Current",
  "published_at": null,
  "modified_at": null,
  "supersedes": [],
  "expires_at": null,
  "formats": {
    "html": "https://feed7.dev/p/2607-05382v1-1xo10v8",
    "json": "https://feed7.dev/p/2607-05382v1-1xo10v8.json",
    "markdown": "https://feed7.dev/p/2607-05382v1-1xo10v8.md"
  }
}