81 lines
2.7 KiB
Python
81 lines
2.7 KiB
Python
import json
|
|
|
|
from ollama import AsyncClient
|
|
from pydantic import BaseModel
|
|
|
|
from haiku.rag.config import Config
|
|
|
|
|
|
class LLMJudgeResponseSchema(BaseModel):
|
|
equivalent: bool
|
|
|
|
|
|
class LLMJudge:
|
|
"""LLM-as-judge for evaluating answer equivalence using Ollama."""
|
|
|
|
def __init__(self, model: str = "qwen3"):
|
|
self.model = model
|
|
self.client = AsyncClient(host=Config.OLLAMA_BASE_URL)
|
|
|
|
async def judge_answers(
|
|
self, question: str, answer: str, expected_answer: str
|
|
) -> bool:
|
|
"""
|
|
Judge whether two answers are equivalent for a given question.
|
|
|
|
Args:
|
|
question: The original question
|
|
answer: The generated answer to evaluate
|
|
expected_answer: The reference/expected answer
|
|
|
|
Returns:
|
|
Dictionary with judgment result:
|
|
- equivalent: bool indicating if answers are equivalent
|
|
- explanation: str explaining the reasoning
|
|
- score: str rating from 1-5
|
|
"""
|
|
|
|
prompt = f"""You are an expert evaluator determining whether two answers to the same question are semantically equivalent.
|
|
|
|
QUESTION: {question}
|
|
|
|
GENERATED ANSWER: {answer}
|
|
|
|
EXPECTED ANSWER: {expected_answer}
|
|
|
|
EVALUATION CRITERIA:
|
|
Rate as EQUIVALENT (true) if:
|
|
✓ Both answers contain the same core factual information
|
|
✓ Both directly address the question asked
|
|
✓ The key claims and conclusions are consistent
|
|
✓ Any additional detail in one answer doesn't contradict the other
|
|
|
|
Rate as NOT EQUIVALENT (false) if:
|
|
✗ Factual contradictions exist between the answers
|
|
✗ One answer fails to address the core question
|
|
✗ Key information is missing from one answer that changes the meaning
|
|
✗ The answers lead to different conclusions or implications
|
|
|
|
GUIDELINES:
|
|
- Ignore minor differences in phrasing, style, or formatting
|
|
- Focus on semantic meaning rather than exact wording
|
|
- Consider both answers correct if they convey the same essential information
|
|
- Be tolerant of different levels of detail if the core answer is preserved
|
|
- Evaluate based on what a person asking this question would need to know
|
|
|
|
Respond with JSON containing only: {{"equivalent": true}} or {{"equivalent": false}}"""
|
|
|
|
response = await self.client.chat(
|
|
model=self.model,
|
|
messages=[{"role": "user", "content": prompt}],
|
|
format=LLMJudgeResponseSchema.model_json_schema(),
|
|
think=False,
|
|
)
|
|
|
|
answer = response["message"]["content"].strip()
|
|
try:
|
|
res = json.loads(answer)
|
|
assert "equivalent" in res, "Response must contain 'equivalent' key"
|
|
return res["equivalent"]
|
|
except json.JSONDecodeError:
|
|
assert False, "Response is not valid JSON"
|