feat: LLM Judge

[monoxgas]

LLM Judge

Key Changes:

• Added a new llm_judge scorer which uses an inference model to grade based on a rubric.

---

Generated Summary:

• Added a new module llm_judge.py to implement a scoring system using a language model (LLM) to evaluate outputs against a rubric.
• Introduced two new classes, JudgeInput and Judgement, to structure the input data and output results respectively.
• Implemented the core function llm_judge, which evaluates the input based on the rubric and provides metrics, including score and pass/fail status.
• Updated __init__.py to include the new llm_judge scorer in the module's exports, ensuring it can be easily accessed.
• Updated pyproject.toml to upgrade the rigging library to version ^3.2.1 for compatibility with the new scoring functionality.
• Potential impact: This feature enhances the scoring mechanism by integrating LLM capabilities, which could improve the accuracy of evaluations based on defined criteria.

This summary was generated with ❤️ by rigging