Настройка уровня лексики

Словарный запас: регулятор удобства чтения

Простой язык без жаргона

import anthropic

client = anthropic.Anthropic(api_key='sk-ant-your-key-here')

response = client.messages.create(
    model='claude-opus-4-5',
    max_tokens=150,
    messages=[{
        'role': 'user',
        'content': (
            'Explain what encryption is. '
            'Simple language only — no technical jargon. '
            'If you must use a technical word, immediately define it in plain English in parentheses. '
            'Target reading level: general public with no tech background. '
            '3 sentences maximum.'
        )
    }]
)
print(response.content[0].text)

Требования к уровню чтения

import openai

client = openai.OpenAI(api_key='sk-your-key-here')

levels = ['5th-grade', '8th-grade', '12th-grade', 'college']

for level in levels:
    response = client.chat.completions.create(
        model='gpt-4o',
        max_tokens=60,
        messages=[{
            'role': 'user',
            'content': (
                f'Explain what a database index is. '
                f'Reading level: {level}. 2 sentences maximum.'
            )
        }]
    )
    print(f'[{level}]: {response.choices[0].message.content.strip()}')
    print()

Ориентация на экспертные знания

import anthropic

client = anthropic.Anthropic(api_key='sk-ant-your-key-here')

expert_prompt = (
    'Explain the trade-off between MVCC and locking-based concurrency control in OLTP databases.\n'
    'Audience: senior database engineers with 10+ years experience.\n'
    'Assume expert knowledge — no definitions of basic terms needed.\n'
    'Use technical terminology freely: transaction isolation levels, write skew, '
    'phantom reads, lock contention, WAL, tuple visibility.\n'
    'Max 4 sentences. Be precise, not introductory.'
)

response = client.messages.create(
    model='claude-opus-4-5',
    max_tokens=200,
    messages=[{'role': 'user', 'content': expert_prompt}]
)
print(response.content[0].text)

Включение технической терминологии

import openai

client = openai.OpenAI(api_key='sk-your-key-here')

response = client.chat.completions.create(
    model='gpt-4o',
    messages=[{
        'role': 'user',
        'content': (
            'Explain how HTTPS works for a developer just starting to learn web security. '
            'Vocabulary strategy:\n'
            '- Introduce technical terms gradually — start simple, add complexity\n'
            '- Bold every new technical term on first use\n'
            '- Immediately follow each bold term with a plain-English definition in parentheses\n'
            '- End with a 4-term Glossary section\n'
            'Length: 150 words max (excluding glossary).'
        )
    }]
)
print(response.choices[0].message.content)

Проверка жаргона

import anthropic

client = anthropic.Anthropic(api_key='sk-ant-your-key-here')

technical_text = (
    'Our platform leverages a microservices architecture with asynchronous event-driven '
    'communication via a Kafka message broker. Each service exposes RESTful endpoints '
    'behind an API gateway that handles rate limiting and JWT authentication. '
    'Horizontal pod autoscaling on Kubernetes ensures elasticity under load.'
)

response = client.messages.create(
    model='claude-opus-4-5',
    max_tokens=300,
    messages=[{
        'role': 'user',
        'content': (
            'Perform a jargon audit on the following paragraph.\n'
            '1. List every technical term a non-technical business stakeholder might not know.\n'
            '2. For each term, provide a plain-English alternative (5 words or fewer).\n'
            '3. Rewrite the paragraph using only the plain-English alternatives.\n\n'
            f'Text:\n{technical_text}'
        )
    }]
)
print(response.content[0].text)

Единообразие словарного запаса

import openai

client = openai.OpenAI(api_key='sk-your-key-here')

response = client.chat.completions.create(
    model='gpt-4o',
    messages=[{
        'role': 'user',
        'content': (
            'Write a 3-section product description for our B2B SaaS platform.\n'
            'VOCABULARY RULES — use these terms consistently throughout:\n'
            '- Use "workspace" not "project", "environment", or "account"\n'
            '- Use "team member" not "user" or "collaborator"\n'
            '- Use "dashboard" not "panel", "console", or "interface"\n'
            '- Use "connect" not "integrate", "link", or "sync"\n'
            'Sections: Overview, Key Benefits, How It Works.\n'
            'Max 100 words total.'
        )
    }]
)
print(response.choices[0].message.content)

Словарный запас для международной аудитории

import anthropic

client = anthropic.Anthropic(api_key='sk-ant-your-key-here')

current_text = (
    'Our platform hits the ground running right out of the box, '
    'giving your team a leg up on the competition. '
    'Once you are up to speed, you will be firing on all cylinders. '
    'The ball is in your court — reach out to kick off your trial.'
)

response = client.messages.create(
    model='claude-opus-4-5',
    max_tokens=200,
    messages=[{
        'role': 'user',
        'content': (
            'Rewrite the following text for an international audience where English is a second language.\n'
            'Rules:\n'
            '- Replace all idioms with literal, plain alternatives\n'
            '- Use only common, short words (prefer "start" over "commence")\n'
            '- No phrasal verbs ("kick off" → "start", "reach out" → "contact")\n'
            '- Active voice only\n\n'
            f'Text:\n{current_text}'
        )
    }]
)
print(response.content[0].text)

Проверка словарного запаса с помощью самопроверки

import openai

client = openai.OpenAI(api_key='sk-your-key-here')

# Step 1: generate with vocabulary constraint
result = client.chat.completions.create(
    model='gpt-4o',
    messages=[{
        'role': 'user',
        'content': 'Explain HTTP status codes in 3 sentences for a non-technical business person. Simple words only.'
    }]
)
content = result.choices[0].message.content
print('GENERATED:')
print(content)
print()

# Step 2: vocabulary self-audit
audit = client.chat.completions.create(
    model='gpt-4o',
    messages=[{
        'role': 'user',
        'content': (
            f'Review the following text. List any words a typical non-technical business person '
            f'(no tech background) might not know. For each flagged word, suggest a simpler alternative.\n\n'
            f'Text:\n{content}'
        )
    }]
)
print('VOCABULARY AUDIT:')
print(audit.choices[0].message.content)

Словарный запас и доверие

import anthropic

client = anthropic.Anthropic(api_key='sk-ant-your-key-here')

# Same topic, calibrated vocabulary for different trust contexts
audiences = [
    ('CRYPTO BEGINNER', 'No technical background, first time reading about cryptocurrency. Plain English, no assumed knowledge.'),
    ('BLOCKCHAIN DEVELOPER', 'Expert: knows EVM, gas fees, consensus mechanisms, smart contract bytecode. No definitions needed.')
]

topic = 'Why Ethereum uses proof-of-stake instead of proof-of-work.'

for label, spec in audiences:
    response = client.messages.create(
        model='claude-opus-4-5',
        max_tokens=100,
        messages=[{
            'role': 'user',
            'content': f'Audience: {spec}\n\nExplain in 2 sentences: {topic}'
        }]
    )
    print(f'[{label}]')
    print(response.content[0].text.strip())
    print()

Словарный запас и голос бренда

import anthropic

client = anthropic.Anthropic(api_key='sk-ant-your-key-here')

brand_vocab_system = (
    'You write content for NorthBridge, a B2B SaaS company. '
    'Brand vocabulary rules:\n'
    '- Say "workspace" not "account" or "project"\n'
    '- Say "connect" not "integrate" or "link"\n'
    '- Say "team" not "users" or "members"\n'
    '- Say "get started" not "onboard" or "begin"\n'
    '- Avoid: "leverage", "synergy", "ecosystem", "seamlessly"'
)

response = client.messages.create(
    model='claude-opus-4-5',
    max_tokens=100,
    system=brand_vocab_system,
    messages=[{
        'role': 'user',
        'content': 'Write a 2-sentence CTA for the homepage inviting users to create an account and start using integrations.'
    }]
)
print(response.content[0].text)

Проверка знаний

Настройка уровня словарного запаса — итоги