Indoxjudge

📊 New Feature: Toxicity Discriminative Metric

1. **Toxicity Evaluation**: We've introduced a new toxicity discriminative metric to help assess the safety and appropriateness of LLM outputs. This feature allows for more comprehensive evaluation of model responses, particularly in scenarios where content moderation is crucial.

2. **Enhanced Safety Analysis**: The new metric complements our existing safety evaluations, providing a more nuanced understanding of potential harmful outputs from LLMs.

🔄 Recap of Recent Improvements

As a reminder, v0.0.6 introduced:

- **Fixed MCDA Score for Safety**: Resolved issues related to the Multi-Criteria Decision Analysis (MCDA) score calculation for safety evaluations.
- **JSON Response Bug Fixes**: Corrected bugs affecting the JSON responses of certain metrics.

And v0.0.4 brought:

- **LLM-Powered Chart Interpretation**: Automated, intelligent analysis of evaluation charts using LLMs.
- **Enhanced Visualization Experience**: Deeper insights into model performance through LLM-powered chart analysis.

⚖️ Why Upgrade?

Upgrading to **IndoxJudge v0.0.8** provides you with:

- New toxicity discriminative metric for more comprehensive safety evaluations
- All the improvements from v0.0.6, including fixed MCDA scoring and enhanced JSON response reliability
- Continued access to powerful features like LLM-powered chart interpretation

These enhancements make IndoxJudge an even more robust and versatile tool for your LLM evaluation needs.

📦 How to Upgrade

Upgrade to the latest version of IndoxJudge using the following command:

bash
pip install --upgrade indoxjudge


We recommend all users upgrade to v0.0.8 to benefit from the new toxicity metric and previous improvements.

🤝 Feedback and Support

Your feedback is crucial in helping us improve IndoxJudge. If you encounter any issues or have suggestions for future enhancements, please don't hesitate to open an issue on our GitHub repository or contact our support team.

Thank you for your continued support and trust in IndoxJudge. We're committed to providing you with the best tools for LLM evaluation and look forward to seeing how these improvements enhance your work.

Happy evaluating!

The IndoxJudge Team

🛠️ Critical Fixes and Improvements

1. **MCDA Score for Safety**: We have addressed and resolved issues related to the Multi-Criteria Decision Analysis (MCDA) score calculation for safety evaluations. This fix ensures more accurate and reliable safety assessments of your LLMs.

2. **JSON Response Bug Fixes**: We've corrected bugs affecting the JSON responses of certain metrics. This improvement enhances the consistency and reliability of the data output, making it easier for users to process and analyze their evaluation results.

🔄 Recap of v0.0.4 Features

As a reminder, v0.0.4 introduced:

- **LLM-Powered Chart Interpretation**: Automated, intelligent analysis of evaluation charts using LLMs.
- **Enhanced Visualization Experience**: Deeper insights into model performance through LLM-powered chart analysis.

⚖️ Why Upgrade?

Upgrading to **IndoxJudge v0.0.6** provides you with:

- Improved accuracy in safety evaluations through fixed MCDA scoring
- Enhanced reliability of JSON responses for various metrics
- All the powerful features from previous versions, including LLM-powered chart interpretation

These improvements make IndoxJudge an even more robust and dependable tool for your LLM evaluation needs.

📦 How to Upgrade

Upgrade to the latest version of IndoxJudge using the following command:

bash
pip install --upgrade indoxjudge


We recommend all users upgrade to v0.0.6 to benefit from these important fixes and improvements.

🤝 Feedback and Support

Your feedback is crucial in helping us improve IndoxJudge. If you encounter any issues or have suggestions for future enhancements, please don't hesitate to open an issue on our GitHub repository or contact our support team.

Thank you for your continued support and trust in IndoxJudge. We're committed to providing you with the best tools for LLM evaluation and look forward to seeing how these improvements enhance your work.

Happy evaluating!

The IndoxJudge Team

🛠️ Critical Fixes and Improvements

1. **MCDA Score for Safety**: We have addressed and resolved issues related to the Multi-Criteria Decision Analysis (MCDA) score calculation for safety evaluations. This fix ensures more accurate and reliable safety assessments of your LLMs.

2. **JSON Response Bug Fixes**: We've corrected bugs affecting the JSON responses of certain metrics. This improvement enhances the consistency and reliability of the data output, making it easier for users to process and analyze their evaluation results.

🔄 Recap of v0.0.4 Features

As a reminder, v0.0.4 introduced:

- **LLM-Powered Chart Interpretation**: Automated, intelligent analysis of evaluation charts using LLMs.
- **Enhanced Visualization Experience**: Deeper insights into model performance through LLM-powered chart analysis.

⚖️ Why Upgrade?

Upgrading to **IndoxJudge v0.0.5** provides you with:

- Improved accuracy in safety evaluations through fixed MCDA scoring
- Enhanced reliability of JSON responses for various metrics
- All the powerful features from previous versions, including LLM-powered chart interpretation

These improvements make IndoxJudge an even more robust and dependable tool for your LLM evaluation needs.

📦 How to Upgrade

Upgrade to the latest version of IndoxJudge using the following command:

bash
pip install --upgrade indoxjudge


We recommend all users upgrade to v0.0.5 to benefit from these important fixes and improvements.

🤝 Feedback and Support

Your feedback is crucial in helping us improve IndoxJudge. If you encounter any issues or have suggestions for future enhancements, please don't hesitate to open an issue on our GitHub repository or contact our support team.

Thank you for your continued support and trust in IndoxJudge. We're committed to providing you with the best tools for LLM evaluation and look forward to seeing how these improvements enhance your work.

Happy evaluating!

The IndoxJudge Team

🧠 LLM-Powered Chart Interpretation

We're excited to introduce a game-changing feature in this release:

**Intelligent Chart Analysis**: Users can now assign an LLM to interpret their evaluation charts automatically. This feature provides insightful analysis of your model's performance, helping you quickly understand trends, patterns, and areas for improvement.

📊 Enhanced Visualization Experience

When users create plots to visualize their evaluation results, they can now leverage the power of LLMs to gain deeper insights into their charts. This feature bridges the gap between raw data and actionable insights, making it easier than ever to understand and act upon your evaluation results.

🔄 Recap of v0.0.3 Features

As a reminder, v0.0.3 introduced:

- **Enhanced RAG Evaluator**: Support for lists of entries for each evaluation metric and the ability to evaluate datasets of questions, responses, and context.
- **Improved Integration with Indox**: Streamlined connection between IndoxJudge and Indox for enhanced workflow efficiency.

⚖️ Why Upgrade?

Upgrading to **IndoxJudge v0.0.4** provides you with:

- Automated, intelligent interpretation of your evaluation charts
- Deeper insights into your model's performance
- Time savings in analysis and decision-making
- All the powerful features from previous versions

Whether you're a researcher, developer, or AI enthusiast, these new features make IndoxJudge an even more indispensable tool in your LLM evaluation toolkit.

📦 How to Upgrade

Upgrade to the latest version of IndoxJudge using the following command:

bash
pip install --upgrade indoxjudge

We're excited to see how you'll use these new features to gain deeper insights into your LLM performance. As always, we welcome your feedback and suggestions for future improvements.

Happy evaluating!

The IndoxJudge Team

📚 **Enhanced RAG Evaluator**
The **Retrieval-Augmented Generation (RAG)** Evaluator has been significantly improved:
- Now supports **lists of entries** for each evaluation metric, streamlining the process of handling complex RAG pipelines.
- Provides the ability to evaluate a **dataset** of questions, responses, and context, ensuring comprehensive evaluation of your RAG models.

🔗 **Improved Integration with Indox**
We’ve made it easier than ever to integrate **IndoxJudge** with **Indox**. This streamlined connection enables seamless collaboration between the two tools, enhancing overall efficiency for users who rely on both platforms.

⚖️ Why Upgrade?
With the latest enhancements in **IndoxJudge v0.0.3**, you gain more robust evaluation capabilities that offer deeper insights into the performance of your models. Whether you are a researcher, developer, or AI enthusiast, these new features make IndoxJudge the perfect tool for refining your LLM evaluation process.

---

📦 How to Upgrade

Upgrade to the latest version of **IndoxJudge** using the following command:

bash
pip install --upgrade indoxjudge

- **🆕 MCDA Scoring for Pipelines:**
Evaluate your models more precisely with our new Multi-Criteria Decision Analysis (MCDA) scoring system! This feature allows you to assign appropriate weights to various metrics, providing a more refined and accurate evaluation score for each pipeline.

- **🛡️ Enhanced Safety Evaluator:**
We've improved our Safety Evaluator by adding new metrics for a more comprehensive analysis. Ensure your LLMs are up to the highest safety standards with greater accuracy and confidence.

- **✨ Robust LLM Evaluation:**
Dive deep into diverse metrics to thoroughly assess your models' performance and capabilities.

- **⚖️ LLM Comparison:**
Easily compare different models side-by-side to determine which one meets your specific needs the best.

- **📚 RAG Evaluator:**
Efficiently evaluate Retrieval-Augmented Generation (RAG) models with our user-friendly interface.

- **📊 Stunning Visualizations:**
Generate beautiful plots and graphs to visualize your results and gain deeper insights into your models' performance.

Why Upgrade?

With these new features, IndoxJudge v0.0.2 makes model evaluation more precise, insightful, and tailored to your specific needs. Whether you're a researcher, developer, or AI enthusiast, this tool is perfect for enhancing your LLM evaluation workflow.

How to Upgrade

Upgrade to the latest version via pip:

bash
pip install --upgrade indoxjudge