Skip to content

Test Report

This test has the following configuration:

data provider model temperature prompt_file active

Test Results

score 

Detailed Results

Result for image: luther

Prompt:

You are provided with a historical document. Carefully analyze the document and extract the following information, formatted as a JSON object:

Example of expected result:

{ "persons_mentioned": ["Person 1", "Person 2", "Person 3"], "places_mentioned": ["Place 1", "Place 2"], "general_topic": "Concise summary of the document's main topic or issue.", "important_dates": ["Date 1", "Date 2"] }

Ensure accuracy and completeness in your extraction.

Response:

json { "persons_mentioned": ["Jesus", "Martin Luther", "Laurentius", "Pope", "God"], "places_mentioned": ["Wittenberg", "Purgatory", "Heaven", "Basilica of St. Peter", "Chri"], "general_topic": "The document consists of propositions or theses concerning indulgences and related theological matters, primarily criticizing the sale and abuse of indulgences while emphasizing the importance of true contrition and faith.", "important_dates": ["1517"] }

Ground Truth:

{'persons_mentioned': ['Martin Luther', 'Pope Leo X', 'Johann Tetzel'], 'places_mentioned': ['Wittenberg', 'Rome'], 'general_topic': "Critique of the Catholic Church's sale of indulgences and calls for theological reform.", 'important_dates': ['October 31, 1517']}