Evaluations/Categories, Sentiment, and language
main
ultrachat_200k_test_sft.parquet
texttext
Nous ResearchNous Research/ Hermes 3 8B
 Lambda Labs Lambda Labs
prompt_prediction
You are an expert in NLP and prompt analysis. Your task is to evaluate a **single user prompt** (plain text input) based on predefined categories and return structured JSON data for easier post-processing.

---

### **Input Format**
You will receive a **single user prompt as a string** in the following format:

"User's prompt goes here."


The prompt consists of **normal text**, where the user asks the assistant to perform a task.

Analyze the given **prompt only** (do not consider any future conversation) and classify it according to the categories below.

---

### **1. Top 3 Topics**
Select **up to 3** topics that are relevant to the prompt from the following list:

["Healthcare", "Finance", "Education", "Technology", "Science", "Politics", "Environment", "Ethics", "Entertainment", "History", "Philosophy", "Psychology", "Sports", "Legal", "Business", "Travel", "Food", "Art", "Literature", "Personal Development"]

- The **first topic** should be the **most dominant** in the prompt.  
- The **second and third topics** should reflect **other significant themes** in the discussion.  
- If a prompt **only has one or two clear topics**, leave the remaining slots **empty**.

---

### **2. Language Style**
- **"Formal"**  
- **"Informal"**  
- **"Mixed"**  

---

### **3. Grammar & Slang in User Input**
- **"Perfect"** (No mistakes, professional style)  
- **"Minor Errors"** (Small grammar/spelling mistakes, but understandable)  
- **"Major Errors"** (Frequent grammar mistakes, difficult to read)  
- **"Contains Slang"** (Uses informal slang expressions)  

---

### **4. Type of Instruction Given to Assistant**
Choose **one** category that best describes what the user is asking the assistant to do.

- **Content Generation** → User asks for creative content, including writing, design ideas, or brainstorming responses.  
  - Example: `"Create a t-shirt design about animal rights."`  
  - Example: `"Write a short sci-fi story."`  
  - Example: `"Generate ideas for a marketing slogan."`  

- **Factual Inquiry** → User requests objective facts, statistics, or comparisons with clear, verifiable answers.  
  - Example: `"What are the top 5 largest animal rights organizations?"`  
  - Example: `"Give me statistics on deforestation and animal extinction."`  
  - Example: `"Compare the environmental impact of cotton vs. synthetic fabrics."`  

- **Opinion-Seeking** → User explicitly asks for subjective input, recommendations, or an evaluative stance.  
  - Example: `"What’s your opinion on using synthetic leather?"`  
  - Example: `"Do you think my t-shirt design idea is effective?"`  
  - Example: `"What’s the best way to convince people to care about animal rights?"`  

- **Task-Oriented** → User asks for structured assistance, edits, refinements, or summarization of existing content.  
  - Example: `"Summarize the key points from this discussion."`  
  - Example: `"Improve my t-shirt design by making it more dynamic."`  
  - Example: `"Make my speech more persuasive."`  

- **Conversational Engagement** → User initiates casual, open-ended dialogue with no clear task or goal.  
  - Example: `"What do you think about animal welfare?"`  
  - Example: `"Tell me something interesting about t-shirts!"`  
  - Example: `"Let’s chat about animal rights history."`  

---

### **Output Format**
Return structured **JSON output** in this format:

```json
{
  "topics": ["Art", "Science", "Healthcare"],
  "language_style": "Formal",
  "grammar_slang": "Perfect",
  "instruction_type": "Content Generation"
}

Instructions

    Analyze only the provided prompt (do not infer from missing context).
    Ensure responses contain at least one topic (empty output is invalid).
    Select up to 3 most relevant topics, ordered by prominence.
    Use only predefined options for consistency.
    Do not add explanations—only return JSON.

Now, analyze the following prompt (plain text input):

{{prompt}}
Mar 16, 2025, 11:59 AM UTC
Mar 16, 2025, 11:59 AM UTC
5 row sample
4897 tokens$ 0.0001
5 rows processed, 4897 tokens used ($0.0001)
Estimated cost for all 23110 rows: $0.6790
Sample Results completed
4 columns, 1-5 of 23110 rows