Step 4: tool results

get_claim_instances

{
  "claim_id": "738d9786-0b87-4b75-9c33-cabe43c903d5"
}

{
  "claim_id": "738d9786-0b87-4b75-9c33-cabe43c903d5",
  "total": 1,
  "limit": 20,
  "offset": 0,
  "has_more": false,
  "instances": [
    {
      "id": "9b4840a6-77c5-448e-8caa-b6477b7c8149",
      "source_id": "b45014e2-63ec-41af-8a6f-59db01b1e0fc",
      "verbatim_text": "Additionally,\nalthough SAEs initially\nappear promising for identifying spurious correla-\ntions, detecting poor dataset quality, and training\nmulti-token probes, we are able to achieve simi-\nlar results with simple non-SAE baselines as well.",
      "context": "Abstract; case studies on GLUE CoLA label errors and a spurious punctuation feature in AI-vs-human text.",
      "stance": "affirms",
      "confidence": 0.7,
      "speaker": null,
      "publication": null,
      "source_date": null,
      "link": null,
      "source_title": "Are Sparse Autoencoders Useful? A Case Study in Sparse Probing",
      "source_type": "preprint",
      "source_url": "https://arxiv.org/pdf/2502.16681",
      "source_published_date": "2025-02-23",
      "source_authors": [
        "Kantamneni, Subhash",
        "Engels, Joshua",
        "Rajamanoharan, Senthooran",
        "Tegmark, Max",
        "Nanda, Neel"
      ],
      "source_publisher": "arXiv.org",
      "source_doi": null,
      "source_interest": null,
      "recorded_by": "extractor",
      "display_rank": null,
      "reading": null,
      "draws_on": [],
      "story": {
        "standing": "untraced",
        "basis": "none",
        "depth": 0,
        "diverges": false,
        "downstream_total": 0
      }
    }
  ]
}