Step 4: tool results

get_claim_instances

{
  "claim_id": "20fee30f-0192-4b8c-b479-83c3f230014d"
}

{
  "claim_id": "20fee30f-0192-4b8c-b479-83c3f230014d",
  "total": 1,
  "limit": 20,
  "offset": 0,
  "has_more": false,
  "instances": [
    {
      "id": "9ec528ab-7fad-40fa-bacc-12c320a2a809",
      "source_id": "447e476f-1376-48e2-9679-30bc4fa0e96c",
      "verbatim_text": "We thought that, for sufficiently good SAEs, a sparse probe in SAE latents would be less likely to overfit to minor spurious correlations compared to a dense probe, and thus that being interpretable gave a valuable inductive bias",
      "context": "The hypothesis motivating the study; the authors later report results contrary to it and express doubt about the argument.",
      "stance": "denies",
      "confidence": 0.75,
      "speaker": null,
      "publication": null,
      "source_date": null,
      "link": null,
      "source_title": "Negative Results for SAEs On Downstream Tasks and Deprioritising SAE Research (GDM Mech Interp Team Progress Update #2)",
      "source_type": "blog_post",
      "source_url": "https://www.lesswrong.com/posts/4uXCAJNuPKtKBsi28/negative-results-for-saes-on-downstream-tasks",
      "source_published_date": null,
      "source_authors": [
        "lewis smith",
        "Senthooran Rajamanoharan",
        "Arthur Conmy",
        "CallumMcDougall",
        "Tom Lieberum",
        "János Kramár",
        "Rohin Shah",
        "Neel Nanda"
      ],
      "source_publisher": null,
      "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
      }
    }
  ]
}