如何使用预构建的评估器

LangSmith 与开源集成openevals包 提供一套预构建的现成计算器，您可以立即将其用作评估的起点。

注意

本操作指南将演示如何设置和运行一种类型的评估员 （LLM-as-a-judge），但还有许多其他可用的评估员。 请参阅 openevals 和 agentevals repos 以获取包含使用示例的完整列表。

设置

您需要安装openevals包来使用预构建的 LLM as-a-judge 评估器。

蟒

pip install -U openevals

TypeScript （类型脚本）

yarn add openevals @langchain/core

您还需要将 OpenAI API 密钥设置为环境变量，但您也可以选择不同的提供商：

export OPENAI_API_KEY="your_openai_api_key"

我们还将使用 LangSmith 的 Python pytest 集成和TypeScript 的 Vitest/Jest 来运行我们的评估。openevals还与evaluate方法。 有关设置说明，请参阅相应的指南。

运行赋值器

一般流程很简单：从openevals，然后在 测试文件。LangSmith 会自动将评估员的结果记录为反馈。

请注意，并非所有赋值器都需要每个参数（例如，完全匹配赋值器只需要 output 和 reference outputs）。 此外，如果您的 LLM-as-a-judge 提示需要其他变量，则将它们作为 kwargs 传入会将它们格式化到提示中。

按如下方式设置测试文件：

蟒

import pytest

from langsmith import testing as t

from openevals.llm import create_llm_as_judge
from openevals.prompts import CORRECTNESS_PROMPT

correctness_evaluator = create_llm_as_judge(
prompt=CORRECTNESS_PROMPT,
feedback_key="correctness",
model="openai:o3-mini",
)

# Mock standin for your application
def my_llm_app(inputs: dict) -> str:
return "Doodads have increased in price by 10% in the past year."

@pytest.mark.langsmith
def test_correctness():
inputs = "How much has the price of doodads changed in the past year?"
reference_outputs = "The price of doodads has decreased by 50% in the past year."
outputs = my_llm_app(inputs)
t.log_inputs({"question": inputs})
t.log_outputs({"answer": outputs})
t.log_reference_outputs({"answer": reference_outputs})

correctness_evaluator(
    inputs=inputs,
    outputs=outputs,
    reference_outputs=reference_outputs
)

TypeScript （类型脚本）

import * as ls from "langsmith/vitest";
// import * as ls from "langsmith/jest";

import { createLLMAsJudge, CORRECTNESS_PROMPT } from "openevals";

const correctnessEvaluator = createLLMAsJudge({
prompt: CORRECTNESS_PROMPT,
feedbackKey: "correctness",
model: "openai:o3-mini",
});

// Mock standin for your application
const myLLMApp = async (_inputs: Record<string, unknown>) => {
return "Doodads have increased in price by 10% in the past year.";
}

ls.describe("Correctness", () => {
ls.test("incorrect answer", {
inputs: {
  question: "How much has the price of doodads changed in the past year?"
},
referenceOutputs: {
  answer: "The price of doodads has decreased by 50% in the past year."
}
}, async ({ inputs, referenceOutputs }) => {
const outputs = await myLLMApp(inputs);
ls.logOutputs({ answer: outputs });

await correctnessEvaluator({
  inputs,
  outputs,
  referenceOutputs,
});
});
});

这feedback_key/feedbackKeyparameter 将用作实验中反馈的名称。

在终端中运行 eval 将导致如下结果：

您还可以将预构建的计算器直接传递到evaluate方法（如果您已在 LangSmith 中创建数据集）。 如果使用 Python，则需要langsmith>=0.3.11:

蟒

from langsmith import Client
from openevals.llm import create_llm_as_judge
from openevals.prompts import CONCISENESS_PROMPT

client = Client()

conciseness_evaluator = create_llm_as_judge(
prompt=CONCISENESS_PROMPT,
feedback_key="conciseness",
model="openai:o3-mini",
)

experiment_results = client.evaluate(
# This is a dummy target function, replace with your actual LLM-based system
lambda inputs: "What color is the sky?",
data="Sample dataset",
evaluators=[
    conciseness_evaluator
]
)

TypeScript （类型脚本）

import { evaluate } from "langsmith/evaluation";
import { createLLMAsJudge, CONCISENESS_PROMPT } from "openevals";

const concisenessEvaluator = createLLMAsJudge({
prompt: CONCISENESS_PROMPT,
feedbackKey: "conciseness",
model: "openai:o3-mini",
});

await evaluate(
(inputs) => "What color is the sky?",
{
data: datasetName,
evaluators: [concisenessEvaluator],
}
);

有关可用评估器的完整列表，请参阅 openevals 和 agentevals 存储库。