| # Metrics | |
| - **sem_div (50%)**: Semantic diversity - cosine distance between sentence embeddings | |
| - **lex_div (50%)**: Lexical diversity - variety of word choice | |
| - **Originality Score**: | |
| # Requirements | |
| - Python 3.8+ | |
| - OpenAI API Key: (https://openai.com/ko-KR/index/openai-api/) | |
| # Run sequence | |
| 1. Download OpenAI_Originality_GUI.py | |
| 2. In Command Prompt, type pip install gradio openai numpy nltk and press Enter. | |
| 3. In Command Prompt, navigate to the folder where OpenAI_Originality_GUI.py is saved. | |
| 4. Type python OpenAI_Originality_GUI.py and press Enter. | |
| 5. A browser opens automatically (or go to the URL shown in the terminal). | |
| 6. (On the Web UI) | |
| - Enter the OpenAI API Key | |
| - Input Passage A | |
| - Input Passage B | |
| 5. Click Submit | |
| # Interpretation | |
| |------------|----------| | |
| | < 5% | Similar | | |
| | 5% - 10% | Some difference | | |
| | 10% - 15% | Significant difference | | |
| | > 15% | Clear difference | | |
| # Model Information | |
| - Model: text-embedding-3-large | |
| - Embedding dimension: 3072 | |
| - Provider: OpenAI | |
| ## License: Apache-2.0 | |