Step 18: tool results

get_claim_instances

{
  "claim_id": "a5665880-9436-4582-a906-e2161fe8e71a"
}

{
  "claim_id": "a5665880-9436-4582-a906-e2161fe8e71a",
  "total": 1,
  "limit": 20,
  "offset": 0,
  "has_more": false,
  "instances": [
    {
      "id": "6912e235-b510-4be2-9a4c-5eb12f2c7841",
      "source_id": "ff9222f2-759d-4782-a7ff-8c811f759780",
      "verbatim_text": "We find that Sparse\nAutoencoders can, in fact, perform close to on par with the reference LoRA performance\non the AxBench benchmark, when features are selected and labelled with our supervised\npipeline.",
      "context": "Abstract; repeated in Results (Figure 2) and Conclusion. Model Gemma-2-9b-it with Gemma Scope SAEs, layers 17 and 32, labels from Stack Exchange tags.",
      "stance": "affirms",
      "confidence": 0.85,
      "speaker": null,
      "publication": null,
      "source_date": null,
      "link": null,
      "source_title": "Steering LLMs? Actually, Sparse Autoencoders can outperform simple baselines",
      "source_type": "preprint",
      "source_url": "https://arxiv.org/pdf/2605.31183",
      "source_published_date": "2026-05-29",
      "source_authors": [
        "Jørgensen, Mikkel Godsk",
        "Hansen, Lars Kai"
      ],
      "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
      }
    }
  ]
}