---
configs:
- config_name: default
  data_files:
  - split: data
    path: "all.jsonl"
---

This dataset consists of LLM-generated items in the style of the Choice of Plausible Alternatives (COPA) benchmark. For more info, see the [paper](https://aclanthology.org/2024.findings-emnlp.299.pdf) and the [Github repo](https://github.com/roemmele/Gen-COPA).