llm-evaluation
Implement comprehensive evaluation strategies for LLM applications using automated metrics, human feedback, and benchmarking. Use when testing LLM performance, measuring AI application quality, or establishing evaluation frameworks.
- License 许可
- MIT
- 最近更新
- 2026-11-01
- Stars
- ★ 3
- Forks
- 1
01 前言 Front Matter
- name
- llm-evaluation
- description
- Implement comprehensive evaluation strategies for LLM applications using automated metrics, human feedback, and benchmarking. Use when testing LLM performance, measuring AI application quality, or establishing evaluation frameworks.
03 SKILL.md 正文
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
from llm_eval import EvaluationSuite, Metric
# Define evaluation suite
suite = EvaluationSuite([
Metric.accuracy(),
Metric.bleu(),
Metric.bertscore(),
Metric.custom(name="groundedness", fn=check_groundedness)
])
# Prepare test cases
test_cases = [
{
"input": "What is the capital of France?",
"expected": "Paris",
"context": "France is a country in Europe. Paris is its capital."
},
# ... more test cases
]
# Run evaluation
results = suite.evaluate(
model=your_model,
test_cases=test_cases
)
print(f"Overall Accuracy: {results.metrics['accuracy']}")
print(f"BLEU Score: {results.metrics['bleu']}")
Automated Metrics Implementation
BLEU Score
from nltk.translate.bleu_score import sentence_bleu, SmoothingFunction
def calculate_bleu(reference, hypothesis):
"""Calculate BLEU score between reference and hypothesis."""
smoothie = SmoothingFunction().method4
return sentence_bleu(
[reference.split()],
hypothesis.split(),
smoothing_function=smoothie
)
# Usage
bleu = calculate_bleu(
reference="The cat sat on the mat",
hypothesis="A cat is sitting on the mat"
)
ROUGE Score
from rouge_score import rouge_scorer
def calculate_rouge(reference, hypothesis):
"""Calculate ROUGE scores."""
scorer = rouge_scorer.RougeScorer(['rouge1', 'rouge2', 'rougeL'], use_stemmer=True)
scores = scorer.score(reference, hypothesis)
return {
'rouge1': scores['rouge1'].fmeasure,
'rouge2': scores['rouge2'].fmeasure,
'rougeL': scores['rougeL'].fmeasure
}
BERTScore
from bert_score import score
def calculate_bertscore(references, hypotheses):
"""Calculate BERTScore using pre-trained BERT."""
P, R, F1 = score(
hypotheses,
references,
lang='en',
model_type='microsoft/deberta-xlarge-mnli'
)
return {
'precision': P.mean().item(),
'recall': R.mean().item(),
'f1': F1.mean().item()
}
Custom Metrics
def calculate_groundedness(response, context):
"""Check if response is grounded in provided context."""
# Use NLI model to check entailment
from transformers import pipeline
nli = pipeline("text-classification", model="microsoft/deberta-large-mnli")
result = nli(f"{context} [SEP] {response}")[0]
# Return confidence that response is entailed by context
return result['score'] if result['label'] == 'ENTAILMENT' else 0.0
def calculate_toxicity(text):
"""Measure toxicity in generated text."""
from detoxify import Detoxify
results = Detoxify('original').predict(text)
return max(results.values()) # Return highest toxicity score
def calculate_factuality(claim, knowledge_base):
"""Verify factual claims against knowledge base."""
# Implementation depends on your knowledge base
# Could use retrieval + NLI, or fact-checking API
pass
LLM-as-Judge Patterns
Single Output Evaluation
def llm_judge_quality(response, question):
"""Use GPT-5 to judge response quality."""
prompt = f"""Rate the following response on a scale of 1-10 for:
1. Accuracy (factually correct)
2. Helpfulness (answers the question)
3. Clarity (well-written and understandable)
Question: {question}
Response: {response}
Provide ratings in JSON format:
{{
"accuracy": <1-10>,
"helpfulness": <1-10>,
"clarity": <1-10>,
"reasoning": "<brief explanation>"
}}
"""
result = openai.ChatCompletion.create(
model="gpt-5",
messages=[{"role": "user", "content": prompt}],
temperature=0
)
return json.loads(result.choices[0].message.content)
Pairwise Comparison
def compare_responses(question, response_a, response_b):
"""Compare two responses using LLM judge."""
prompt = f"""Compare these two responses to the question and determine which is better.
Question: {question}
Response A: {response_a}
Response B: {response_b}
Which response is better and why? Consider accuracy, helpfulness, and clarity.
Answer with JSON:
{{
"winner": "A" or "B" or "tie",
"reasoning": "<explanation>",
"confidence": <1-10>
}}
"""
result = openai.ChatCompletion.create(
model="gpt-5",
messages=[{"role": "user", "content": prompt}],
temperature=0
)
return json.loads(result.choices[0].message.content)
Human Evaluation Frameworks
Annotation Guidelines
class AnnotationTask:
"""Structure for human annotation task."""
def __init__(self, response, question, context=None):
self.response = response
self.question = question
self.context = context
def get_annotation_form(self):
return {
"question": self.question,
"context": self.context,
"response": self.response,
"ratings": {
"accuracy": {
"scale": "1-5",
"description": "Is the response factually correct?"
},
"relevance": {
"scale": "1-5",
"description": "Does it answer the question?"
},
"coherence": {
"scale": "1-5",
"description": "Is it logically consistent?"
}
},
"issues": {
"factual_error": False,
"hallucination": False,
"off_topic": False,
"unsafe_content": False
},
"feedback": ""
}
Inter-Rater Agreement
from sklearn.metrics import cohen_kappa_score
def calculate_agreement(rater1_scores, rater2_scores):
"""Calculate inter-rater agreement."""
kappa = cohen_kappa_score(rater1_scores, rater2_scores)
interpretation = {
kappa < 0: "Poor",
kappa < 0.2: "Slight",
kappa < 0.4: "Fair",
kappa < 0.6: "Moderate",
kappa < 0.8: "Substantial",
kappa <= 1.0: "Almost Perfect"
}
return {
"kappa": kappa,
"interpretation": interpretation[True]
}
A/B Testing
Statistical Testing Framework
from scipy import stats
import numpy as np
class ABTest:
def __init__(self, variant_a_name="A", variant_b_name="B"):
self.variant_a = {"name": variant_a_name, "scores": []}
self.variant_b = {"name": variant_b_name, "scores": []}
def add_result(self, variant, score):
"""Add evaluation result for a variant."""
if variant == "A":
self.variant_a["scores"].append(score)
else:
self.variant_b["scores"].append(score)
def analyze(self, alpha=0.05):
"""Perform statistical analysis."""
a_scores = self.variant_a["scores"]
b_scores = self.variant_b["scores"]
# T-test
t_stat, p_value = stats.ttest_ind(a_scores, b_scores)
# Effect size (Cohen's d)
pooled_std = np.sqrt((np.std(a_scores)**2 + np.std(b_scores)**2) / 2)
cohens_d = (np.mean(b_scores) - np.mean(a_scores)) / pooled_std
return {
"variant_a_mean": np.mean(a_scores),
"variant_b_mean": np.mean(b_scores),
"difference": np.mean(b_scores) - np.mean(a_scores),
"relative_improvement": (np.mean(b_scores) - np.mean(a_scores)) / np.mean(a_scores),
"p_value": p_value,
"statistically_significant": p_value < alpha,
"cohens_d": cohens_d,
"effect_size": self.interpret_cohens_d(cohens_d),
"winner": "B" if np.mean(b_scores) > np.mean(a_scores) else "A"
}
@staticmethod
def interpret_cohens_d(d):
"""Interpret Cohen's d effect size."""
abs_d = abs(d)
if abs_d < 0.2:
return "negligible"
elif abs_d < 0.5:
return "small"
elif abs_d < 0.8:
return "medium"
else:
return "large"
Regression Testing
Regression Detection
class RegressionDetector:
def __init__(self, baseline_results, threshold=0.05):
self.baseline = baseline_results
self.threshold = threshold
def check_for_regression(self, new_results):
"""Detect if new results show regression."""
regressions = []
for metric in self.baseline.keys():
baseline_score = self.baseline[metric]
new_score = new_results.get(metric)
if new_score is None:
continue
# Calculate relative change
relative_change = (new_score - baseline_score) / baseline_score
# Flag if significant decrease
if relative_change < -self.threshold:
regressions.append({
"metric": metric,
"baseline": baseline_score,
"current": new_score,
"change": relative_change
})
return {
"has_regression": len(regressions) > 0,
"regressions": regressions
}
Benchmarking
Running Benchmarks
class BenchmarkRunner:
def __init__(self, benchmark_dataset):
self.dataset = benchmark_dataset
def run_benchmark(self, model, metrics):
"""Run model on benchmark and calculate metrics."""
results = {metric.name: [] for metric in metrics}
for example in self.dataset:
# Generate prediction
prediction = model.predict(example["input"])
# Calculate each metric
for metric in metrics:
score = metric.calculate(
prediction=prediction,
reference=example["reference"],
context=example.get("context")
)
results[metric.name].append(score)
# Aggregate results
return {
metric: {
"mean": np.mean(scores),
"std": np.std(scores),
"min": min(scores),
"max": max(scores)
}
for metric, scores in results.items()
}
Resources
- references/metrics.md: Comprehensive metric guide
- references/human-evaluation.md: Annotation best practices
- references/benchmarking.md: Standard benchmarks
- references/a-b-testing.md: Statistical testing guide
- references/regression-testing.md: CI/CD integration
- assets/evaluation-framework.py: Complete evaluation harness
- assets/benchmark-dataset.jsonl: Example datasets
- scripts/evaluate-model.py: Automated evaluation runner
Best Practices
- Multiple Metrics: Use diverse metrics for comprehensive view
- Representative Data: Test on real-world, diverse examples
- Baselines: Always compare against baseline performance
- Statistical Rigor: Use proper statistical tests for comparisons
- Continuous Evaluation: Integrate into CI/CD pipeline
- Human Validation: Combine automated metrics with human judgment
- Error Analysis: Investigate failures to understand weaknesses
- Version Control: Track evaluation results over time
Common Pitfalls
- Single Metric Obsession: Optimizing for one metric at the expense of others
- Small Sample Size: Drawing conclusions from too few examples
- Data Contamination: Testing on training data
- Ignoring Variance: Not accounting for statistical uncertainty
- Metric Mismatch: Using metrics not aligned with business goals
04 同仓库更多 Skills
seedance-prompt-en
Write effective prompts for Jimeng Seedance 2.0 multimodal AI video generation. Use when users want to create video prompts using text, images, videos, and audio inputs with the @ reference system. Covers camera movements, effects replication, video extension, editing, music beat-matching, e-commerce ads, short dramas, and educational content.
onepage-pdf
Convert an HTML page into a single continuous-page PDF (one tall page, no pagination breaks), preserving desktop layout, backgrounds and selectable text. Use when the user asks to turn an HTML file/report/proposal into a "single-page PDF", "long PDF", "不分页 PDF", "单页 PDF", "长图式 PDF", or complains that an HTML-to-PDF export breaks into pages or cuts content. Supports optional string redaction with leak verification.
native-feel-cross-platform-desktop
Use when the user is designing, prototyping, or rewriting a desktop app that must run on multiple OSes (macOS + Windows, optionally Linux) AND feel indistinguishable from a native app to its users — fast launch, native windowing, native input handling, native materials. Trigger words include "cross-platform desktop", "Electron alternative", "Tauri vs native", "WebView wrapper", "near-native performance", "Raycast architecture", "WebKit/WebView2 quirks", "WKWebView", "system tray app", "global hotkey app", "launcher app". Do NOT trigger this skill for pure web apps, pure mobile apps, or for greenfield projects that have no native-feel requirement.
gstack
Router for the gstack skill suite. (gstack)
fireworks-tech-graph
Use when the user wants to create any technical diagram - architecture, data flow, flowchart, sequence, agent/memory, or concept map - and export as SVG+PNG. Trigger on: "画图" "帮我画" "生成图" "做个图" "架构图" "流程图" "可视化一下" "出图" "generate diagram" "draw diagram" "visualize" or any system/flow description the user wants illustrated.
dbs-video-extract
接收抖音、小红书、微信视频号等短视频链接或分享文案,查询作品/账号数据并提取语音文字稿,按作者和标题归档为 Markdown。用户希望完整解析短视频、取得数据或文字稿时使用;运行前检查 TikHub 与轻抖使用凭证,缺少时用新手能理解的方式引导购买和安全配置。