wshobson/agents

llm-evaluation

Implement comprehensive evaluation strategies for LLM applications using automated metrics, human feedback, and benchmarking.

소스 보기

사용 방법

  1. 설치 명령을 복사하고 자신의 에이전트 환경에 설치하기 전에 소스 패키지를 확인하세요.
  2. 필요한 도구, 인증 정보, 지원 환경은 원본 문서에서 확인하세요.
  3. 입력과 원하는 산출물을 설명하고 출처와 결과를 대조하세요.
npx skills add https://github.com/wshobson/agents --skill llm-evaluation

시작 전 확인

  • 설치에는 모델 구독, API 한도, 파일 접근 권한이 포함되지 않습니다.
  • 별점 수와 설치 수는 활동 지표이며 독립적인 품질·보안 검증이 아닙니다.
원본 Skill 문서

원본 저장소의 제목, 예시, 코드, 표, 링크, 이미지를 유지해 표시합니다.

LLM Evaluation

Master comprehensive evaluation strategies for LLM applications, from automated metrics to human evaluation and A/B testing.

When to Use This Skill

  • Measuring LLM application performance systematically
  • Comparing different models or prompts
  • Detecting performance regressions before deployment
  • Validating improvements from prompt changes
  • Building confidence in production systems
  • Establishing baselines and tracking progress over time
  • Debugging unexpected model behavior

Core Evaluation Types

1. Automated Metrics

Fast, repeatable, scalable evaluation using computed scores.

Text Generation:

  • BLEU: N-gram overlap (translation)
  • ROUGE: Recall-oriented (summarization)
  • METEOR: Semantic similarity
  • BERTScore: Embedding-based similarity
  • Perplexity: Language model confidence

Classification:

  • Accuracy: Percentage correct
  • Precision/Recall/F1: Class-specific performance
  • Confusion Matrix: Error patterns
  • AUC-ROC: Ranking quality

Retrieval (RAG):

  • MRR: Mean Reciprocal Rank
  • NDCG: Normalized Discounted Cumulative Gain
  • Precision@K: Relevant in top K
  • Recall@K: Coverage in top K

2. Human Evaluation

Manual assessment for quality aspects difficult to automate.

Dimensions:

  • Accuracy: Factual correctness
  • Coherence: Logical flow
  • Relevance: Answers the question
  • Fluency: Natural language quality
  • Safety: No harmful content
  • Helpfulness: Useful to the user

3. LLM-as-Judge

Use stronger LLMs to evaluate weaker model outputs.

Approaches:

  • Pointwise: Score individual responses
  • Pairwise: Compare two responses
  • Reference-based: Compare to gold standard
  • Reference-free: Judge without ground truth

Quick Start

python
from dataclasses import dataclass
from typing import Callable
import numpy as np

@dataclass
class Metric:
    name: str
    fn: Callable

    @staticmethod
    def accuracy():
        return Metric("accuracy", calculate_accuracy)

    @staticmethod
    def bleu():
        return Metric("bleu", calculate_bleu)

    @staticmethod
    def bertscore():
        return Metric("bertscore", calculate_bertscore)

    @staticmethod
    def custom(name: str, fn: Callable):
        return Metric(name, fn)

class EvaluationSuite:
    def __init__(self, metrics: list[Metric]):
        self.metrics = metrics

    async def evaluate(self, model, test_cases: list[dict]) -> dict:
        results = {m.name: [] for m in self.metrics}

        for test in test_cases:
            prediction = await model.predict(test["input"])

            for metric in self.metrics:
                score = metric.fn(
                    prediction=prediction,
                    reference=test.get("expected"),
                    context=test.get("context")
                )
                results[metric.name].append(score)

        return {
            "metrics": {k: np.mean(v) for k, v in results.items()},
            "raw_scores": results
        }

# Usage
suite = EvaluationSuite([
    Metric.accuracy(),
    Metric.bleu(),
    Metric.bertscore(),
    Metric.custom("groundedness", check_groundedness)
])

test_cases = [
    {
        "input": "What is the capital of France?",
        "expected": "Paris",
        "context": "France is a country in Europe. Paris is its capital."
    },
]

results = await suite.evaluate(model=your_model, test_cases=test_cases)

Detailed patterns and worked examples

Detailed pattern documentation lives in references/details.md. Read that file when the navigation tier above is insufficient.

관련 작업

같은 저장소의 Skills

더 많은 Skills

모든 Skills
wshobson
커뮤니티

typescript-advanced-types

Master TypeScript's advanced type system including generics, conditional types, mapped types, template literals, and utility types for building type-safe applications. Use when implementing complex type logic, creating reusable type utilities, or ensuring compile-time type safety in TypeScript projects.

누적 설치 수
10.1만
GitHub Stars
4.3만
업데이트
11월 21일
wshobson
커뮤니티

tailwind-design-system

Build scalable design systems with Tailwind CSS v4, design tokens, component libraries, and responsive patterns. Use when creating component libraries, implementing design systems, or standardizing UI patterns.

누적 설치 수
7.5만
GitHub Stars
4.3만
업데이트
11월 18일
wshobson
커뮤니티

kpi-dashboard-design

Design effective KPI dashboards with metrics selection, visualization best practices, and real-time monitoring patterns. Use this skill when building an executive SaaS metrics dashboard tracking MRR, churn, and LTV/CAC ratios; designing an operations center with live service health and request throughput; creating a cohort retention analysis view for a product team; or debugging a dashboard where metrics contradict each other due to inconsistent calculation methodology.

1시간 설치 수
2
GitHub Stars
4.3만
업데이트
11월 18일
wshobson
커뮤니티

market-sizing-analysis

Calculate TAM/SAM/SOM for market opportunities using top-down, bottom-up, and value theory methodologies. Use this skill when sizing markets, estimating addressable revenue, validating market opportunity for a new venture, or building investor-ready market analysis for a startup pitch or business plan.

1시간 설치 수
2
GitHub Stars
4.3만
업데이트
11월 18일