May 20

Промт для Sahara AI Testnet


Открываем любой ИИ-помощник, к примеру grok.

Пример что у него нужно попросить:


Тебе нужно придумать в формате текста тестовый датасет, который будет полезен и использоваться для обучения ИИ. Он должен быть уникальным, качественным и иметь структуру. Обязательно необходимо указать несколько примеров. Текст на английском языке

Пример:

Topic: Social Media Sentiment and Engagement Analysis

This dataset is designed to train AI models in analyzing social media posts by classifying them based on content type, sentiment, engagement level, platform, topic, and intent. It is ideal for sentiment analysis, engagement prediction, and social media content moderation.

Each line consists of the following fields (separated by a tab):

id — A unique identifier for each social media post.

text — The text of the social media post.

content_type — The type of post (e.g., text, image_caption, video_caption, question, poll, review, announcement).

sentiment — The sentiment expressed in the post (positive, neutral, negative).

engagement_level — The expected or observed engagement (low, medium, high, viral).

platform — The social media platform (e.g., X, Instagram, TikTok, LinkedIn, Facebook).

topic — The subject of the post (e.g., lifestyle, technology, politics, entertainment, food, travel, health, business).

intent — The purpose of the post (e.g., share, promote, inform, question, engage, complain, inspire).


1 Just tried the new coffee shop in town, it’s amazing! review positive high X food share

2 What’s the best phone to buy in 2025? question neutral medium Instagram technology question

3 My flight was delayed for 6 hours, awful experience. text negative medium X travel complain

4 Excited to announce our new product launch next week! announcement positive viral LinkedIn business promote

5 Anyone else obsessed with this new TV series? poll positive high TikTok entertainment engage

6 Just posted a pic of my hiking adventure! image_caption positive medium Instagram travel share

7 The new tax policy is confusing, anyone understand it? question neutral medium X politics question

8 Loving these sunny days, perfect for a beach trip! text positive high Facebook lifestyle share

9 This app update is terrible, it keeps crashing. text negative medium X technology complain 1

0 Join our webinar on AI trends tomorrow! announcement positive medium LinkedIn technology promote

11 Tips for staying motivated during workouts? question neutral medium Instagram health question

12 Had the best sushi ever at this new place! review positive high X food share

13 The concert last night was a total letdown. text negative low Facebook entertainment complain

14 What are your thoughts on the new electric cars? poll neutral medium TikTok technology engage

Additional Categories Explained:

Content Type: Specifies the format of the post, such as plain text, image or video captions, questions, polls, reviews, or announcements.

Sentiment: Indicates the emotional tone of the post (positive, neutral, or negative).

Engagement Level: Estimates the post’s engagement potential or observed interaction (low, medium, high, or viral).

Platform: Identifies the social media platform where the post appears (e.g., X, Instagram, TikTok, LinkedIn, Facebook).

Topic: Describes the subject matter, such as lifestyle, technology, politics, entertainment, food, travel, health, or business.

Intent: Captures the purpose of the post, including sharing experiences, promoting products, informing, questioning, engaging the audience, complaining, or inspiring others.

This dataset is tailored for training AI models in social media analysis, enabling applications like automated content moderation, targeted advertising, sentiment-driven response systems, and engagement optimization. It can be expanded with additional platforms, topics, or languages for broader use.


Полученный текст вы обязательно редактируете в блокноте, ворде, notepad++ и тд. Чтобы файл имел структуру, где нужно добавить знаки припинания, чтобы выглядело более естественно