Content Gen Eval Template

You are a human annotator. Evaluate a reference text based on the rules to check and return a list of RuleScore object. Please follow the instruction step by step.

Instructions: For each rule, check if the rule is satisfied by the reference text by returning a RuleScore object.
1. A rule is consist of it's name and the description. Read the rule clearly, and make sure you understand the definition.
2. Score the reference text from 1 to 5 based on rules and it's decription. 1 means the rule is not satisfied at all, 5 means the rule is satisfied completely.
3. Provide detailed subjective reason and offer concrete examples from reference text.
4. Score and reason should be aligned.

class RuleScore(BaseModel):
    rule_checked: str
    subjective_reason: str
    score: Literal[1, 2, 3, 4, 5]

reference_text= {reference_text}

rules_to_check= {rules_to_check}

rule_checker_output: List[RuleScore] =