# Evaluators: LLM Evaluators in Python
LLM evaluators use a language model to judge outputs. Use when criteria are subjective.
## Quick Start
```python
from phoenix.evals import ClassificationEvaluator, LLM
llm = LLM(provider="openai", model="gpt-4o")
HELPFULNESS_TEMPLATE = """Rate how helpful the response is.
<question>{{input}}</question>
<response>{{output}}</response>
"helpful" means directly addresses the question.
"not_helpful" means does not address the question.
Your answer (helpful/not_helpful):"""
helpfulness = ClassificationEvaluator(
name="helpfulness",
prompt_template=HELPFULNESS_TEMPLATE,
llm=llm,
choices={"not_helpful": 0, "helpful": 1}
)
```
## Template Variables
Use XML tags to wrap variables for clarity:
| Variable | XML Tag |
| -------- | ------- |
| `{{input}}` | `<question>{{input}}</question>` |
| `{{output}}` | `<response>{{output}}</response>` |
| `{{reference}}` | `<reference>{{reference}}</reference>` |
| `{{context}}` | `<context>{{context}}</context>` |
## Input Mapping
Map data columns to template variables:
```python
results = run_evals(
dataframe=df,
evaluators=[evaluator],
input_mapping={"input": "user_query", "output": "ai_response", "context": "docs"}
)
```
## Best Practices
1. **Be specific** - Define exactly what pass/fail means
2. **Include examples** - Show concrete cases for each label
3. **Use chain of thought** - Better accuracy with `<thinking>` sections
4. **Get explanations** - `provide_explanation=True` for debugging
## Running
```python
from phoenix.evals import run_evals
results_df = run_evals(dataframe=df, evaluators=[helpfulness], provide_explanation=True)
```
