feat: add reference geometry stability benchmark

[lwehrhan]

Pre-review checklist for PR author

PR author must check the checkboxes below when creating the PR.

• I've confirmed the contribution guidelines.
• I have reviewed and understand all AI-generated code in this PR.
• I have added human-written tests for the new logic.
• I have properly cited any upstream algorithms or libraries the AI utilized.
• I have disclosed significant AI tool usage in the PR description.

Summary

This benchmark assesses the MLIP's capability to preserve the ground-state geometry of organic small molecules during energy minimization. It draws its test set from the OpenFF industry dataset, which contains 73,301 conformers of thousands of different molecules, collected by industry partners of the OpenFF consortium. 200 neutral molecules and 20 charged molecules were picked from this dataset, maximizing chemical diversity based on Morgan fingerprints.

Linked issue

Resolves #664

Progress

• Calculations
• Analysis
• Application
• Documentation

Testing

New decorators/callbacks