# Beyond Scores: Understanding LLM-as-a-Judge Mechanisms in Summarization Evaluation

This repository accompanies the paper *Beyond Scores: Understanding LLM-as-a-Judge Mechanisms in Summarization Evaluation*, accepted to the EMNLP 2026 Main Conference.

## Status

The full release, including the perturbation taxonomy, generation pipeline, mechanistic analysis scripts, and behaviorally-verified corpus is currently in preparation and will be released here in due course.

## Citation

Full citation information will be added once the paper appears in the ACL Anthology.

## Contact

For questions, contact vasava [at] wisc [dot] edu.