RISE Humanities Data Benchmark, 0.5.0-pre1

Search Test Runs

 

A test run is a single execution of a benchmark test using a defined model configuration.
Each run represents how a particular large language model (LLM) — such as GPT-4, Claude-3, or Gemini — performed on a given task at a specific time, with specific settings.

A test run includes:

  • Prompt and role definition – what the model was asked to do and from what perspective (e.g. “as a historian”).
  • Model configuration – provider, model version, temperature, and other generation parameters.
  • Results – the model’s actual response and its evaluation (scores such as F1 or accuracy).
  • Usage and cost data – token counts and calculated API costs.
  • Metadata – information like the test date, benchmark name, and person who executed it.

Together, test runs make it possible to compare models, providers, and configurations across benchmarks in a transparent and reproducible way.

Search Results
Show compact results Refine Search New Search
Download JSON Download CSV

Your search for Benchmark 'library_cards__true' with Search Hidden 'False' returned 105 results, showing page 7 of 11.
Result 61 of 105

Test T0164 at 2025-10-03

{'document-type': ['index-card'], 'writing': ['typed', 'printed', 'handwritten'], 'century': [20, 19], 'language': ['de', 'fr', 'en', 'la', 'el', 'fi', 'sv', 'pl'], 'layout': ['index'], 'entry-type': ['bibliographic'], 'task': ['information-extraction']}

Configuration
Provideropenai
Modelgpt-4o-mini
  
Temperature0.0
DataclassDocument
  
Normalized Score0.95 %
Test timeunknown seconds
Prompt 

Extract the bibliographic information about a historical dissertation from this index card and return it as a structured JSON object with the following exact format:

```json
{
  "type": {
    "type": "Dissertation or thesis" OR "Reference"
  },
  "author": {
    "last_name": "string",
    "first_name": "string"
  },
  "publication": {
    "title": "string",
    "year": integer,
    "place": "string or empty string",
    "pages": "string or empty string",
    "publisher": "string or empty string",
    "format": "string or empty string"
  },
  "library_reference": {
    "shelfmark": "string or empty string",
    "subjects": "string or empty string"
  }
}
```

EXTRACTION RULES:
1. **Card Type**: If a card contains the note "s." on a separate line, it is a "Reference". Otherwise, it is a "Dissertation or thesis".

2. **Author**: Extract last_name and first_name. If only one name is given, put it in last_name and leave first_name empty.

3. **Publication**:
   - title: The main title of the work
   - year: Publication year as integer
   - place: Publication place
   - pages: Page count (remove " S." suffix if present)
   - publisher: Publishing house/institution
   - format: Usually "8°", "8'", or "4°" - single value only

4. **Library Reference**:
   - shelfmark: Often begins with "Diss." or "AT", may be marked with "Standort:"
   - subjects: Subject classifications or keywords

5. **Missing Information**: Use empty string "" for missing text fields, omit year fields entirely if not present.

6. **Ignore**: Disregard any information that doesn't fit into these categories.

Return ONLY the JSON object, no additional text or explanation.

Results 

no valid result

Scoring
Fuzzy Score F1 micro / macro Micro precision/recall Tue/False Positives
n/a 0.01 0.01 0.07 0.01 263 20 275 2395
      Micro Precision Micro Recall Instances TP FP FN
Costs / Pricing
Pricing Date: 6 months ago, 2025-10-03. Tokens: 131.3K IT + 727 OT = 132.0K TTCost: 0.020$ + 0.000$ = 0.020$
Cite: Hindermann, Marti, Alberto, et al., (2025). RISE-UNIBAS/humanities_data_benchmark, 10.5281/zenodo.16941752
Result 62 of 105

Test T0162 at 2025-10-02

{'document-type': ['index-card'], 'writing': ['typed', 'printed', 'handwritten'], 'century': [20, 19], 'language': ['de', 'fr', 'en', 'la', 'el', 'fi', 'sv', 'pl'], 'layout': ['index'], 'entry-type': ['bibliographic'], 'task': ['information-extraction']}

Configuration
Provideropenai
Modelgpt-4.1-nano
  
Temperature0.0
DataclassDocument
  
Normalized Score66.70 %
Test timeunknown seconds
Prompt 

Extract the bibliographic information about a historical dissertation from this index card and return it as a structured JSON object with the following exact format:

```json
{
  "type": {
    "type": "Dissertation or thesis" OR "Reference"
  },
  "author": {
    "last_name": "string",
    "first_name": "string"
  },
  "publication": {
    "title": "string",
    "year": integer,
    "place": "string or empty string",
    "pages": "string or empty string",
    "publisher": "string or empty string",
    "format": "string or empty string"
  },
  "library_reference": {
    "shelfmark": "string or empty string",
    "subjects": "string or empty string"
  }
}
```

EXTRACTION RULES:
1. **Card Type**: If a card contains the note "s." on a separate line, it is a "Reference". Otherwise, it is a "Dissertation or thesis".

2. **Author**: Extract last_name and first_name. If only one name is given, put it in last_name and leave first_name empty.

3. **Publication**:
   - title: The main title of the work
   - year: Publication year as integer
   - place: Publication place
   - pages: Page count (remove " S." suffix if present)
   - publisher: Publishing house/institution
   - format: Usually "8°", "8'", or "4°" - single value only

4. **Library Reference**:
   - shelfmark: Often begins with "Diss." or "AT", may be marked with "Standort:"
   - subjects: Subject classifications or keywords

5. **Missing Information**: Use empty string "" for missing text fields, omit year fields entirely if not present.

6. **Ignore**: Disregard any information that doesn't fit into these categories.

Return ONLY the JSON object, no additional text or explanation.

Results 

no valid result

Scoring
Fuzzy Score F1 micro / macro Micro precision/recall Tue/False Positives
n/a 0.68 0.67 0.72 0.64 263 1536 592 879
      Micro Precision Micro Recall Instances TP FP FN
Costs / Pricing
Pricing Date: None, None. Tokens: None IT + None OT = None TTCost: None$ + None$ = None$
Cite: Hindermann, Marti, Alberto, et al., (2025). RISE-UNIBAS/humanities_data_benchmark, 10.5281/zenodo.16941752
Result 63 of 105

Test T0161 at 2025-10-02

{'document-type': ['index-card'], 'writing': ['typed', 'printed', 'handwritten'], 'century': [20, 19], 'language': ['de', 'fr', 'en', 'la', 'el', 'fi', 'sv', 'pl'], 'layout': ['index'], 'entry-type': ['bibliographic'], 'task': ['information-extraction']}

Configuration
Provideropenai
Modelgpt-4.1-mini
  
Temperature0.0
DataclassDocument
  
Normalized Score75.74 %
Test timeunknown seconds
Prompt 

Extract the bibliographic information about a historical dissertation from this index card and return it as a structured JSON object with the following exact format:

```json
{
  "type": {
    "type": "Dissertation or thesis" OR "Reference"
  },
  "author": {
    "last_name": "string",
    "first_name": "string"
  },
  "publication": {
    "title": "string",
    "year": integer,
    "place": "string or empty string",
    "pages": "string or empty string",
    "publisher": "string or empty string",
    "format": "string or empty string"
  },
  "library_reference": {
    "shelfmark": "string or empty string",
    "subjects": "string or empty string"
  }
}
```

EXTRACTION RULES:
1. **Card Type**: If a card contains the note "s." on a separate line, it is a "Reference". Otherwise, it is a "Dissertation or thesis".

2. **Author**: Extract last_name and first_name. If only one name is given, put it in last_name and leave first_name empty.

3. **Publication**:
   - title: The main title of the work
   - year: Publication year as integer
   - place: Publication place
   - pages: Page count (remove " S." suffix if present)
   - publisher: Publishing house/institution
   - format: Usually "8°", "8'", or "4°" - single value only

4. **Library Reference**:
   - shelfmark: Often begins with "Diss." or "AT", may be marked with "Standort:"
   - subjects: Subject classifications or keywords

5. **Missing Information**: Use empty string "" for missing text fields, omit year fields entirely if not present.

6. **Ignore**: Disregard any information that doesn't fit into these categories.

Return ONLY the JSON object, no additional text or explanation.

Results 

no valid result

Scoring
Fuzzy Score F1 micro / macro Micro precision/recall Tue/False Positives
n/a 0.76 0.76 0.76 0.76 263 1834 584 581
      Micro Precision Micro Recall Instances TP FP FN
Costs / Pricing
Pricing Date: None, None. Tokens: None IT + None OT = None TTCost: None$ + None$ = None$
Cite: Hindermann, Marti, Alberto, et al., (2025). RISE-UNIBAS/humanities_data_benchmark, 10.5281/zenodo.16941752
Result 64 of 105

Test T0208 at 2025-10-01

{'document-type': ['index-card'], 'writing': ['typed', 'printed', 'handwritten'], 'century': [20, 19], 'language': ['de', 'fr', 'en', 'la', 'el', 'fi', 'sv', 'pl'], 'layout': ['index'], 'entry-type': ['bibliographic'], 'task': ['information-extraction']}

Configuration
Providergenai
Modelgemini-2.5-flash-lite
  
Temperature0.0
DataclassDocument
  
Normalized Score68.92 %
Test timeunknown seconds
Prompt 

Extract the bibliographic information about a historical dissertation from this index card and return it as a structured JSON object with the following exact format:

```json
{
  "type": {
    "type": "Dissertation or thesis" OR "Reference"
  },
  "author": {
    "last_name": "string",
    "first_name": "string"
  },
  "publication": {
    "title": "string",
    "year": integer,
    "place": "string or empty string",
    "pages": "string or empty string",
    "publisher": "string or empty string",
    "format": "string or empty string"
  },
  "library_reference": {
    "shelfmark": "string or empty string",
    "subjects": "string or empty string"
  }
}
```

EXTRACTION RULES:
1. **Card Type**: If a card contains the note "s." on a separate line, it is a "Reference". Otherwise, it is a "Dissertation or thesis".

2. **Author**: Extract last_name and first_name. If only one name is given, put it in last_name and leave first_name empty.

3. **Publication**:
   - title: The main title of the work
   - year: Publication year as integer
   - place: Publication place
   - pages: Page count (remove " S." suffix if present)
   - publisher: Publishing house/institution
   - format: Usually "8°", "8'", or "4°" - single value only

4. **Library Reference**:
   - shelfmark: Often begins with "Diss." or "AT", may be marked with "Standort:"
   - subjects: Subject classifications or keywords

5. **Missing Information**: Use empty string "" for missing text fields, omit year fields entirely if not present.

6. **Ignore**: Disregard any information that doesn't fit into these categories.

Return ONLY the JSON object, no additional text or explanation.

Results 

no valid result

Scoring
Fuzzy Score F1 micro / macro Micro precision/recall Tue/False Positives
n/a 0.70 0.69 0.71 0.68 263 1648 663 767
      Micro Precision Micro Recall Instances TP FP FN
Costs / Pricing
Pricing Date: 6 months ago, 2025-10-01. Tokens: 185.4K IT + 214.7K OT = 400.1K TTCost: 0.019$ + 0.086$ = 0.104$
Cite: Hindermann, Marti, Alberto, et al., (2025). RISE-UNIBAS/humanities_data_benchmark, 10.5281/zenodo.16941752
Result 65 of 105

Test T0151 at 2025-10-01

{'document-type': ['index-card'], 'writing': ['typed', 'printed', 'handwritten'], 'century': [20, 19], 'language': ['de', 'fr', 'en', 'la', 'el', 'fi', 'sv', 'pl'], 'layout': ['index'], 'entry-type': ['bibliographic'], 'task': ['information-extraction']}

Configuration
Providergenai
Modelgemini-2.0-flash
  
Temperature0.0
DataclassDocument
  
Normalized Score83.23 %
Test timeunknown seconds
Prompt 

Extract the bibliographic information about a historical dissertation from this index card and return it as a structured JSON object with the following exact format:

```json
{
  "type": {
    "type": "Dissertation or thesis" OR "Reference"
  },
  "author": {
    "last_name": "string",
    "first_name": "string"
  },
  "publication": {
    "title": "string",
    "year": integer,
    "place": "string or empty string",
    "pages": "string or empty string",
    "publisher": "string or empty string",
    "format": "string or empty string"
  },
  "library_reference": {
    "shelfmark": "string or empty string",
    "subjects": "string or empty string"
  }
}
```

EXTRACTION RULES:
1. **Card Type**: If a card contains the note "s." on a separate line, it is a "Reference". Otherwise, it is a "Dissertation or thesis".

2. **Author**: Extract last_name and first_name. If only one name is given, put it in last_name and leave first_name empty.

3. **Publication**:
   - title: The main title of the work
   - year: Publication year as integer
   - place: Publication place
   - pages: Page count (remove " S." suffix if present)
   - publisher: Publishing house/institution
   - format: Usually "8°", "8'", or "4°" - single value only

4. **Library Reference**:
   - shelfmark: Often begins with "Diss." or "AT", may be marked with "Standort:"
   - subjects: Subject classifications or keywords

5. **Missing Information**: Use empty string "" for missing text fields, omit year fields entirely if not present.

6. **Ignore**: Disregard any information that doesn't fit into these categories.

Return ONLY the JSON object, no additional text or explanation.

Results 

no valid result

Scoring
Fuzzy Score F1 micro / macro Micro precision/recall Tue/False Positives
n/a 0.84 0.83 0.85 0.84 263 2023 370 392
      Micro Precision Micro Recall Instances TP FP FN
Costs / Pricing
Pricing Date: 6 months ago, 2025-10-01. Tokens: 612.5K IT + 45.7K OT = 658.2K TTCost: 0.061$ + 0.018$ = 0.080$
Cite: Hindermann, Marti, Alberto, et al., (2025). RISE-UNIBAS/humanities_data_benchmark, 10.5281/zenodo.16941752
Result 66 of 105

Test T0165 at 2025-10-01

{'document-type': ['index-card'], 'writing': ['typed', 'printed', 'handwritten'], 'century': [20, 19], 'language': ['de', 'fr', 'en', 'la', 'el', 'fi', 'sv', 'pl'], 'layout': ['index'], 'entry-type': ['bibliographic'], 'task': ['information-extraction']}

Configuration
Provideropenai
Modelgpt-5
  
Temperature0.0
DataclassDocument
  
Normalized Score86.27 %
Test timeunknown seconds
Prompt 

Extract the bibliographic information about a historical dissertation from this index card and return it as a structured JSON object with the following exact format:

```json
{
  "type": {
    "type": "Dissertation or thesis" OR "Reference"
  },
  "author": {
    "last_name": "string",
    "first_name": "string"
  },
  "publication": {
    "title": "string",
    "year": integer,
    "place": "string or empty string",
    "pages": "string or empty string",
    "publisher": "string or empty string",
    "format": "string or empty string"
  },
  "library_reference": {
    "shelfmark": "string or empty string",
    "subjects": "string or empty string"
  }
}
```

EXTRACTION RULES:
1. **Card Type**: If a card contains the note "s." on a separate line, it is a "Reference". Otherwise, it is a "Dissertation or thesis".

2. **Author**: Extract last_name and first_name. If only one name is given, put it in last_name and leave first_name empty.

3. **Publication**:
   - title: The main title of the work
   - year: Publication year as integer
   - place: Publication place
   - pages: Page count (remove " S." suffix if present)
   - publisher: Publishing house/institution
   - format: Usually "8°", "8'", or "4°" - single value only

4. **Library Reference**:
   - shelfmark: Often begins with "Diss." or "AT", may be marked with "Standort:"
   - subjects: Subject classifications or keywords

5. **Missing Information**: Use empty string "" for missing text fields, omit year fields entirely if not present.

6. **Ignore**: Disregard any information that doesn't fit into these categories.

Return ONLY the JSON object, no additional text or explanation.

Results 

no valid result

Scoring
Fuzzy Score F1 micro / macro Micro precision/recall Tue/False Positives
n/a 0.87 0.86 0.87 0.87 263 2095 305 320
      Micro Precision Micro Recall Instances TP FP FN
Costs / Pricing
Pricing Date: 6 months ago, 2025-10-01. Tokens: 364.8K IT + 666.8K OT = 1.0M TTCost: 0.456$ + 6.668$ = 7.124$
Cite: Hindermann, Marti, Alberto, et al., (2025). RISE-UNIBAS/humanities_data_benchmark, 10.5281/zenodo.16941752
Result 67 of 105

Test T0180 at 2025-10-01

{'document-type': ['index-card'], 'writing': ['typed', 'printed', 'handwritten'], 'century': [20, 19], 'language': ['de', 'fr', 'en', 'la', 'el', 'fi', 'sv', 'pl'], 'layout': ['index'], 'entry-type': ['bibliographic'], 'task': ['information-extraction']}

Configuration
Providermistral
Modelmistral-medium-2505
  
Temperature0.0
DataclassDocument
  
Normalized Score77.35 %
Test timeunknown seconds
Prompt 

Extract the bibliographic information about a historical dissertation from this index card and return it as a structured JSON object with the following exact format:

```json
{
  "type": {
    "type": "Dissertation or thesis" OR "Reference"
  },
  "author": {
    "last_name": "string",
    "first_name": "string"
  },
  "publication": {
    "title": "string",
    "year": integer,
    "place": "string or empty string",
    "pages": "string or empty string",
    "publisher": "string or empty string",
    "format": "string or empty string"
  },
  "library_reference": {
    "shelfmark": "string or empty string",
    "subjects": "string or empty string"
  }
}
```

EXTRACTION RULES:
1. **Card Type**: If a card contains the note "s." on a separate line, it is a "Reference". Otherwise, it is a "Dissertation or thesis".

2. **Author**: Extract last_name and first_name. If only one name is given, put it in last_name and leave first_name empty.

3. **Publication**:
   - title: The main title of the work
   - year: Publication year as integer
   - place: Publication place
   - pages: Page count (remove " S." suffix if present)
   - publisher: Publishing house/institution
   - format: Usually "8°", "8'", or "4°" - single value only

4. **Library Reference**:
   - shelfmark: Often begins with "Diss." or "AT", may be marked with "Standort:"
   - subjects: Subject classifications or keywords

5. **Missing Information**: Use empty string "" for missing text fields, omit year fields entirely if not present.

6. **Ignore**: Disregard any information that doesn't fit into these categories.

Return ONLY the JSON object, no additional text or explanation.

Results 

no valid result

Scoring
Fuzzy Score F1 micro / macro Micro precision/recall Tue/False Positives
n/a 0.78 0.77 0.79 0.76 263 1847 500 568
      Micro Precision Micro Recall Instances TP FP FN
Costs / Pricing
Pricing Date: 6 months ago, 2025-10-01. Tokens: 462.2K IT + 41.3K OT = 503.5K TTCost: 0.185$ + 0.083$ = 0.267$
Cite: Hindermann, Marti, Alberto, et al., (2025). RISE-UNIBAS/humanities_data_benchmark, 10.5281/zenodo.16941752
Result 68 of 105

Test T0146 at 2025-10-01

{'document-type': ['index-card'], 'writing': ['typed', 'printed', 'handwritten'], 'century': [20, 19], 'language': ['de', 'fr', 'en', 'la', 'el', 'fi', 'sv', 'pl'], 'layout': ['index'], 'entry-type': ['bibliographic'], 'task': ['information-extraction']}

Configuration
Provideranthropic
Modelclaude-opus-4-1-20250805
  
Temperature0.0
DataclassDocument
  
Normalized Score81.54 %
Test timeunknown seconds
Prompt 

Extract the bibliographic information about a historical dissertation from this index card and return it as a structured JSON object with the following exact format:

```json
{
  "type": {
    "type": "Dissertation or thesis" OR "Reference"
  },
  "author": {
    "last_name": "string",
    "first_name": "string"
  },
  "publication": {
    "title": "string",
    "year": integer,
    "place": "string or empty string",
    "pages": "string or empty string",
    "publisher": "string or empty string",
    "format": "string or empty string"
  },
  "library_reference": {
    "shelfmark": "string or empty string",
    "subjects": "string or empty string"
  }
}
```

EXTRACTION RULES:
1. **Card Type**: If a card contains the note "s." on a separate line, it is a "Reference". Otherwise, it is a "Dissertation or thesis".

2. **Author**: Extract last_name and first_name. If only one name is given, put it in last_name and leave first_name empty.

3. **Publication**:
   - title: The main title of the work
   - year: Publication year as integer
   - place: Publication place
   - pages: Page count (remove " S." suffix if present)
   - publisher: Publishing house/institution
   - format: Usually "8°", "8'", or "4°" - single value only

4. **Library Reference**:
   - shelfmark: Often begins with "Diss." or "AT", may be marked with "Standort:"
   - subjects: Subject classifications or keywords

5. **Missing Information**: Use empty string "" for missing text fields, omit year fields entirely if not present.

6. **Ignore**: Disregard any information that doesn't fit into these categories.

Return ONLY the JSON object, no additional text or explanation.

Results 

no valid result

Scoring
Fuzzy Score F1 micro / macro Micro precision/recall Tue/False Positives
n/a 0.83 0.82 0.82 0.83 263 2014 448 401
      Micro Precision Micro Recall Instances TP FP FN
Costs / Pricing
Pricing Date: 6 months ago, 2025-10-01. Tokens: 566.6K IT + 54.3K OT = 620.8K TTCost: 8.499$ + 4.069$ = 12.568$
Cite: Hindermann, Marti, Alberto, et al., (2025). RISE-UNIBAS/humanities_data_benchmark, 10.5281/zenodo.16941752
Result 69 of 105

Test T0167 at 2025-10-01

{'document-type': ['index-card'], 'writing': ['typed', 'printed', 'handwritten'], 'century': [20, 19], 'language': ['de', 'fr', 'en', 'la', 'el', 'fi', 'sv', 'pl'], 'layout': ['index'], 'entry-type': ['bibliographic'], 'task': ['information-extraction']}

Configuration
Provideropenai
Modelgpt-5-nano
  
Temperature0.0
DataclassDocument
  
Normalized Score76.59 %
Test timeunknown seconds
Prompt 

Extract the bibliographic information about a historical dissertation from this index card and return it as a structured JSON object with the following exact format:

```json
{
  "type": {
    "type": "Dissertation or thesis" OR "Reference"
  },
  "author": {
    "last_name": "string",
    "first_name": "string"
  },
  "publication": {
    "title": "string",
    "year": integer,
    "place": "string or empty string",
    "pages": "string or empty string",
    "publisher": "string or empty string",
    "format": "string or empty string"
  },
  "library_reference": {
    "shelfmark": "string or empty string",
    "subjects": "string or empty string"
  }
}
```

EXTRACTION RULES:
1. **Card Type**: If a card contains the note "s." on a separate line, it is a "Reference". Otherwise, it is a "Dissertation or thesis".

2. **Author**: Extract last_name and first_name. If only one name is given, put it in last_name and leave first_name empty.

3. **Publication**:
   - title: The main title of the work
   - year: Publication year as integer
   - place: Publication place
   - pages: Page count (remove " S." suffix if present)
   - publisher: Publishing house/institution
   - format: Usually "8°", "8'", or "4°" - single value only

4. **Library Reference**:
   - shelfmark: Often begins with "Diss." or "AT", may be marked with "Standort:"
   - subjects: Subject classifications or keywords

5. **Missing Information**: Use empty string "" for missing text fields, omit year fields entirely if not present.

6. **Ignore**: Disregard any information that doesn't fit into these categories.

Return ONLY the JSON object, no additional text or explanation.

Results 

no valid result

Scoring
Fuzzy Score F1 micro / macro Micro precision/recall Tue/False Positives
n/a 0.77 0.77 0.76 0.78 263 1887 596 528
      Micro Precision Micro Recall Instances TP FP FN
Costs / Pricing
Pricing Date: 6 months ago, 2025-10-01. Tokens: 434.6K IT + 807.7K OT = 1.2M TTCost: 0.022$ + 0.323$ = 0.345$
Cite: Hindermann, Marti, Alberto, et al., (2025). RISE-UNIBAS/humanities_data_benchmark, 10.5281/zenodo.16941752
Result 70 of 105

Test T0166 at 2025-10-01

{'document-type': ['index-card'], 'writing': ['typed', 'printed', 'handwritten'], 'century': [20, 19], 'language': ['de', 'fr', 'en', 'la', 'el', 'fi', 'sv', 'pl'], 'layout': ['index'], 'entry-type': ['bibliographic'], 'task': ['information-extraction']}

Configuration
Provideropenai
Modelgpt-5-mini
  
Temperature0.0
DataclassDocument
  
Normalized Score78.94 %
Test timeunknown seconds
Prompt 

Extract the bibliographic information about a historical dissertation from this index card and return it as a structured JSON object with the following exact format:

```json
{
  "type": {
    "type": "Dissertation or thesis" OR "Reference"
  },
  "author": {
    "last_name": "string",
    "first_name": "string"
  },
  "publication": {
    "title": "string",
    "year": integer,
    "place": "string or empty string",
    "pages": "string or empty string",
    "publisher": "string or empty string",
    "format": "string or empty string"
  },
  "library_reference": {
    "shelfmark": "string or empty string",
    "subjects": "string or empty string"
  }
}
```

EXTRACTION RULES:
1. **Card Type**: If a card contains the note "s." on a separate line, it is a "Reference". Otherwise, it is a "Dissertation or thesis".

2. **Author**: Extract last_name and first_name. If only one name is given, put it in last_name and leave first_name empty.

3. **Publication**:
   - title: The main title of the work
   - year: Publication year as integer
   - place: Publication place
   - pages: Page count (remove " S." suffix if present)
   - publisher: Publishing house/institution
   - format: Usually "8°", "8'", or "4°" - single value only

4. **Library Reference**:
   - shelfmark: Often begins with "Diss." or "AT", may be marked with "Standort:"
   - subjects: Subject classifications or keywords

5. **Missing Information**: Use empty string "" for missing text fields, omit year fields entirely if not present.

6. **Ignore**: Disregard any information that doesn't fit into these categories.

Return ONLY the JSON object, no additional text or explanation.

Results 

no valid result

Scoring
Fuzzy Score F1 micro / macro Micro precision/recall Tue/False Positives
n/a 0.79 0.79 0.77 0.81 263 1959 579 456
      Micro Precision Micro Recall Instances TP FP FN
Costs / Pricing
Pricing Date: 6 months ago, 2025-10-01. Tokens: 387.5K IT + 609.5K OT = 997.0K TTCost: 0.097$ + 1.219$ = 1.316$
Cite: Hindermann, Marti, Alberto, et al., (2025). RISE-UNIBAS/humanities_data_benchmark, 10.5281/zenodo.16941752