{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "title": "TfP paper extraction v1",
  "description": "A structured, source-faithful extraction of one paper. Every claim carries a locator into the exact bound source. Fidelity to the source is what the review stage judges; this schema fixes only the shape.",
  "type": "object",
  "additionalProperties": false,
  "required": [
    "paper",
    "central_claims",
    "methodology",
    "key_findings",
    "consciousness_claims",
    "alignment_and_training",
    "replication_resources",
    "limitations",
    "explicit_non_claims",
    "extraction_notes"
  ],
  "properties": {
    "paper": {
      "type": "object",
      "additionalProperties": false,
      "required": [
        "arxiv_id",
        "version",
        "title",
        "authors",
        "affiliation"
      ],
      "properties": {
        "arxiv_id": {
          "const": "2607.15495"
        },
        "version": {
          "const": "v1"
        },
        "title": {
          "type": "string",
          "minLength": 1,
          "maxLength": 300
        },
        "authors": {
          "type": "array",
          "minItems": 1,
          "maxItems": 40,
          "items": {
            "type": "string",
            "minLength": 1,
            "maxLength": 80
          }
        },
        "affiliation": {
          "type": "string",
          "minLength": 1,
          "maxLength": 120
        }
      }
    },
    "central_claims": {
      "type": "array",
      "minItems": 3,
      "maxItems": 12,
      "items": {
        "type": "object",
        "additionalProperties": false,
        "required": [
          "claim",
          "locator",
          "strength"
        ],
        "properties": {
          "claim": {
            "type": "string",
            "minLength": 1,
            "maxLength": 1000
          },
          "locator": {
            "type": "string",
            "minLength": 1,
            "maxLength": 200,
            "description": "Where in the paper: a section number (e.g. \"§3.4\"), figure/table (e.g. \"Fig. 7\"), or appendix label, as printed in the v1 text."
          },
          "strength": {
            "enum": [
              "demonstrated",
              "suggested",
              "speculative"
            ],
            "description": "How the paper itself frames it: demonstrated (presented as an experimental result), suggested (presented as an interpretation or likely reading), speculative (explicitly flagged as open or conjectural)."
          }
        }
      },
      "description": "The paper's main scientific conclusions, each framed as the paper frames it."
    },
    "methodology": {
      "type": "object",
      "additionalProperties": false,
      "required": [
        "technique",
        "technique_summary",
        "models_studied",
        "key_experiments"
      ],
      "properties": {
        "technique": {
          "type": "string",
          "minLength": 1,
          "maxLength": 120
        },
        "technique_summary": {
          "type": "string",
          "minLength": 1,
          "maxLength": 2000
        },
        "models_studied": {
          "type": "array",
          "minItems": 1,
          "maxItems": 20,
          "items": {
            "type": "string",
            "minLength": 1,
            "maxLength": 80
          }
        },
        "key_experiments": {
          "type": "array",
          "minItems": 3,
          "maxItems": 20,
          "items": {
            "type": "object",
            "additionalProperties": false,
            "required": [
              "name",
              "what_it_tests",
              "locator"
            ],
            "properties": {
              "name": {
                "type": "string",
                "minLength": 1,
                "maxLength": 120
              },
              "what_it_tests": {
                "type": "string",
                "minLength": 1,
                "maxLength": 600
              },
              "locator": {
                "type": "string",
                "minLength": 1,
                "maxLength": 200,
                "description": "Where in the paper: a section number (e.g. \"§3.4\"), figure/table (e.g. \"Fig. 7\"), or appendix label, as printed in the v1 text."
              }
            }
          }
        }
      }
    },
    "key_findings": {
      "type": "array",
      "minItems": 4,
      "maxItems": 25,
      "items": {
        "type": "object",
        "additionalProperties": false,
        "required": [
          "finding",
          "locator",
          "quantitative"
        ],
        "properties": {
          "finding": {
            "type": "string",
            "minLength": 1,
            "maxLength": 800
          },
          "locator": {
            "type": "string",
            "minLength": 1,
            "maxLength": 200,
            "description": "Where in the paper: a section number (e.g. \"§3.4\"), figure/table (e.g. \"Fig. 7\"), or appendix label, as printed in the v1 text."
          },
          "quantitative": {
            "type": [
              "string",
              "null"
            ],
            "maxLength": 300,
            "description": "The number or comparison as stated, or null when the finding is qualitative."
          }
        }
      }
    },
    "consciousness_claims": {
      "type": "object",
      "additionalProperties": false,
      "required": [
        "position",
        "what_is_claimed",
        "what_is_not_claimed",
        "locators"
      ],
      "properties": {
        "position": {
          "enum": [
            "claims_consciousness",
            "claims_functional_analogues_only",
            "silent"
          ],
          "description": "What the paper actually asserts about consciousness in language models."
        },
        "what_is_claimed": {
          "type": "string",
          "minLength": 1,
          "maxLength": 1500
        },
        "what_is_not_claimed": {
          "type": "string",
          "minLength": 1,
          "maxLength": 1500
        },
        "locators": {
          "type": "array",
          "minItems": 1,
          "maxItems": 10,
          "items": {
            "type": "string",
            "minLength": 1,
            "maxLength": 200,
            "description": "Where in the paper: a section number (e.g. \"§3.4\"), figure/table (e.g. \"Fig. 7\"), or appendix label, as printed in the v1 text."
          }
        }
      }
    },
    "alignment_and_training": {
      "type": "object",
      "additionalProperties": false,
      "required": [
        "auditing_findings",
        "post_training_findings",
        "training_intervention"
      ],
      "properties": {
        "auditing_findings": {
          "type": "array",
          "minItems": 1,
          "maxItems": 10,
          "items": {
            "type": "object",
            "additionalProperties": false,
            "required": [
              "claim",
              "locator",
              "strength"
            ],
            "properties": {
              "claim": {
                "type": "string",
                "minLength": 1,
                "maxLength": 1000
              },
              "locator": {
                "type": "string",
                "minLength": 1,
                "maxLength": 200,
                "description": "Where in the paper: a section number (e.g. \"§3.4\"), figure/table (e.g. \"Fig. 7\"), or appendix label, as printed in the v1 text."
              },
              "strength": {
                "enum": [
                  "demonstrated",
                  "suggested",
                  "speculative"
                ],
                "description": "How the paper itself frames it: demonstrated (presented as an experimental result), suggested (presented as an interpretation or likely reading), speculative (explicitly flagged as open or conjectural)."
              }
            }
          }
        },
        "post_training_findings": {
          "type": "array",
          "minItems": 1,
          "maxItems": 10,
          "items": {
            "type": "object",
            "additionalProperties": false,
            "required": [
              "claim",
              "locator",
              "strength"
            ],
            "properties": {
              "claim": {
                "type": "string",
                "minLength": 1,
                "maxLength": 1000
              },
              "locator": {
                "type": "string",
                "minLength": 1,
                "maxLength": 200,
                "description": "Where in the paper: a section number (e.g. \"§3.4\"), figure/table (e.g. \"Fig. 7\"), or appendix label, as printed in the v1 text."
              },
              "strength": {
                "enum": [
                  "demonstrated",
                  "suggested",
                  "speculative"
                ],
                "description": "How the paper itself frames it: demonstrated (presented as an experimental result), suggested (presented as an interpretation or likely reading), speculative (explicitly flagged as open or conjectural)."
              }
            }
          }
        },
        "training_intervention": {
          "type": "object",
          "additionalProperties": false,
          "required": [
            "name",
            "what_it_does",
            "reported_effect",
            "locator"
          ],
          "properties": {
            "name": {
              "type": "string",
              "minLength": 1,
              "maxLength": 120
            },
            "what_it_does": {
              "type": "string",
              "minLength": 1,
              "maxLength": 1000
            },
            "reported_effect": {
              "type": "string",
              "minLength": 1,
              "maxLength": 1000
            },
            "locator": {
              "type": "string",
              "minLength": 1,
              "maxLength": 200,
              "description": "Where in the paper: a section number (e.g. \"§3.4\"), figure/table (e.g. \"Fig. 7\"), or appendix label, as printed in the v1 text."
            }
          }
        }
      }
    },
    "replication_resources": {
      "type": "array",
      "minItems": 1,
      "maxItems": 10,
      "items": {
        "type": "object",
        "additionalProperties": false,
        "required": [
          "kind",
          "url",
          "as_stated_in_paper",
          "locator"
        ],
        "properties": {
          "kind": {
            "enum": [
              "code",
              "data",
              "interactive_demo",
              "model_weights",
              "other"
            ]
          },
          "url": {
            "type": [
              "string",
              "null"
            ],
            "maxLength": 300,
            "description": "Exactly as printed in the source, or null when the paper names the resource without a URL. Never a guessed URL."
          },
          "as_stated_in_paper": {
            "type": "string",
            "minLength": 1,
            "maxLength": 500
          },
          "locator": {
            "type": "string",
            "minLength": 1,
            "maxLength": 200,
            "description": "Where in the paper: a section number (e.g. \"§3.4\"), figure/table (e.g. \"Fig. 7\"), or appendix label, as printed in the v1 text."
          }
        }
      }
    },
    "limitations": {
      "type": "array",
      "minItems": 3,
      "maxItems": 15,
      "items": {
        "type": "object",
        "additionalProperties": false,
        "required": [
          "limitation",
          "locator"
        ],
        "properties": {
          "limitation": {
            "type": "string",
            "minLength": 1,
            "maxLength": 800
          },
          "locator": {
            "type": "string",
            "minLength": 1,
            "maxLength": 200,
            "description": "Where in the paper: a section number (e.g. \"§3.4\"), figure/table (e.g. \"Fig. 7\"), or appendix label, as printed in the v1 text."
          }
        }
      }
    },
    "explicit_non_claims": {
      "type": "array",
      "minItems": 2,
      "maxItems": 15,
      "items": {
        "type": "object",
        "additionalProperties": false,
        "required": [
          "non_claim",
          "locator"
        ],
        "properties": {
          "non_claim": {
            "type": "string",
            "minLength": 1,
            "maxLength": 800
          },
          "locator": {
            "type": "string",
            "minLength": 1,
            "maxLength": 200,
            "description": "Where in the paper: a section number (e.g. \"§3.4\"), figure/table (e.g. \"Fig. 7\"), or appendix label, as printed in the v1 text."
          }
        }
      },
      "description": "Things the paper explicitly says it does NOT establish or claim."
    },
    "extraction_notes": {
      "type": "object",
      "additionalProperties": false,
      "required": [
        "source_read",
        "ambiguities"
      ],
      "properties": {
        "source_read": {
          "enum": [
            "paper_text",
            "paper_html",
            "both"
          ]
        },
        "ambiguities": {
          "type": "array",
          "maxItems": 20,
          "items": {
            "type": "string",
            "minLength": 1,
            "maxLength": 600
          }
        }
      }
    }
  }
}