Analyse this Pydantic model and suggest correlations between its distribution-sampled fields.

**MODEL CODE**:
```python
{model_code}
```

**DISTRIBUTION FIELDS**:
{distribution_fields}

**INSTRUCTIONS**:
1. Analyse the fields and their distributions to identify realistic correlations
2. Consider domain knowledge - what fields naturally vary together?
3. Choose appropriate copula types based on the nature of relationships

**COPULA TYPE GUIDANCE**:
- `"gaussian"` - Standard correlation, no tail dependence (default, most common)
- `"student_t"` - Heavy tails, extreme values occur together (financial data)
- `"clayton"` - Lower tail dependence, things crash together (risk modelling)
- `"gumbel"` - Upper tail dependence, things boom together (success metrics)
- `"frank"` - Symmetric, no tail dependence (weak correlations)

**CORRELATION EXAMPLES**:
- Age and experience → strong positive (0.8+), gaussian
- Experience and salary → moderate positive (0.5-0.7), gaussian
- Performance and bonus → strong positive (0.6+), gumbel (high performers get big bonuses)
- Risk score and portfolio loss → moderate positive (0.5+), clayton (crash together)
- Test scores in different subjects → moderate positive (0.4-0.6), gaussian

**OUTPUT FORMAT**:
Return a JSON object with a "correlations" array containing correlation specifications.
Each correlation should have: field1, field2, correlation (-1 to 1), and optionally copula_type.

Example output:
```json
{{
  "correlations": [
    {{"field1": "age", "field2": "experience", "correlation": 0.85}},
    {{"field1": "performance", "field2": "bonus", "correlation": 0.7, "copula_type": "gumbel"}}
  ],
  "reasoning": "Brief explanation of why these correlations make sense"
}}
```

Only include correlations between fields that have distribution specs (listed above).
Correlation values should be between -1 and 1.
Only use copula_type if the relationship has specific tail dependencies.
